Starbucks_Capstone_Project.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.3"
    },
    "colab": {
      "name": "Starbucks_Capstone_notebook.ipynb",
      "provenance": []
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4wmOvZB7yBHR",
        "colab_type": "text"
      },
      "source": [
        "# Starbucks Capstone Challenge\n",
        "\n",
        "### Introduction\n",
        "\n",
        "This data set contains simulated data that mimics customer behavior on the Starbucks rewards mobile app. Once every few days, Starbucks sends out an offer to users of the mobile app. An offer can be merely an advertisement for a drink or an actual offer such as a discount or BOGO (buy one get one free). Some users might not receive any offer during certain weeks. \n",
        "\n",
        "Not all users receive the same offer, and that is the challenge to solve with this data set.\n",
        "\n",
        "Your task is to combine transaction, demographic and offer data to determine which demographic groups respond best to which offer type. This data set is a simplified version of the real Starbucks app because the underlying simulator only has one product whereas Starbucks actually sells dozens of products.\n",
        "\n",
        "Every offer has a validity period before the offer expires. As an example, a BOGO offer might be valid for only 5 days. You'll see in the data set that informational offers have a validity period even though these ads are merely providing information about a product; for example, if an informational offer has 7 days of validity, you can assume the customer is feeling the influence of the offer for 7 days after receiving the advertisement.\n",
        "\n",
        "You'll be given transactional data showing user purchases made on the app including the timestamp of purchase and the amount of money spent on a purchase. This transactional data also has a record for each offer that a user receives as well as a record for when a user actually views the offer. There are also records for when a user completes an offer. \n",
        "\n",
        "Keep in mind as well that someone using the app might make a purchase through the app without having received an offer or seen an offer.\n",
        "\n",
        "### Example\n",
        "\n",
        "To give an example, a user could receive a discount offer buy 10 dollars get 2 off on Monday. The offer is valid for 10 days from receipt. If the customer accumulates at least 10 dollars in purchases during the validity period, the customer completes the offer.\n",
        "\n",
        "However, there are a few things to watch out for in this data set. Customers do not opt into the offers that they receive; in other words, a user can receive an offer, never actually view the offer, and still complete the offer. For example, a user might receive the \"buy 10 dollars get 2 dollars off offer\", but the user never opens the offer during the 10 day validity period. The customer spends 15 dollars during those ten days. There will be an offer completion record in the data set; however, the customer was not influenced by the offer because the customer never viewed the offer.\n",
        "\n",
        "### Cleaning\n",
        "\n",
        "This makes data cleaning especially important and tricky.\n",
        "\n",
        "You'll also want to take into account that some demographic groups will make purchases even if they don't receive an offer. From a business perspective, if a customer is going to make a 10 dollar purchase without an offer anyway, you wouldn't want to send a buy 10 dollars get 2 dollars off offer. You'll want to try to assess what a certain demographic group will buy when not receiving any offers.\n",
        "\n",
        "### Final Advice\n",
        "\n",
        "Because this is a capstone project, you are free to analyze the data any way you see fit. For example, you could build a machine learning model that predicts how much someone will spend based on demographics and offer type. Or you could build a model that predicts whether or not someone will respond to an offer. Or, you don't need to build a machine learning model at all. You could develop a set of heuristics that determine what offer you should send to each customer (i.e., 75 percent of women customers who were 35 years old responded to offer A vs 40 percent from the same demographic to offer B, so send offer A)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vfQ4Bun4yBHS",
        "colab_type": "text"
      },
      "source": [
        "# Project Goal\n",
        "\n",
        "Based on the context above, this project will try to ask the questions below\n",
        "\n",
        "- What factors mainly affect the usage of the offer from customer? Should the company send out the offer or not?\n",
        "- How possible will a customer open and use the offer sent to them? Is there any common characteristics on the customers who take the offer?\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TIbclFY4yBHT",
        "colab_type": "text"
      },
      "source": [
        "## Implementation\n",
        "In order to answer the questions above, given the data sets provided, the following steps will seperate the data set into three types of offer and fit the data into three supervised classification models. Using this way, the model can predict whether the offer will be reponded by customer or not when sent to them. Also, by investing the feature importance of the model, it can aswer the question that what factors mainly affect people to make the decision and finally comlete the transaction.\n",
        "\n",
        "In order to find out the best models, in this project, I'll initiate the simple decision tree classifier model as a baseline model. And I'll also use random forest as an comparison to baseline model to see if I could get a better accuracy."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9MXEAarSyBHU",
        "colab_type": "text"
      },
      "source": [
        "# Data Sets\n",
        "\n",
        "The data is contained in three files:\n",
        "\n",
        "* portfolio.json - containing offer ids and meta data about each offer (duration, type, etc.)\n",
        "* profile.json - demographic data for each customer\n",
        "* transcript.json - records for transactions, offers received, offers viewed, and offers completed\n",
        "\n",
        "Here is the schema and explanation of each variable in the files:\n",
        "\n",
        "**portfolio.json**\n",
        "* id (string) - offer id\n",
        "* offer_type (string) - type of offer ie BOGO, discount, informational\n",
        "* difficulty (int) - minimum required spend to complete an offer\n",
        "* reward (int) - reward given for completing an offer\n",
        "* duration (int) - time for offer to be open, in days\n",
        "* channels (list of strings)\n",
        "\n",
        "**profile.json**\n",
        "* age (int) - age of the customer \n",
        "* became_member_on (int) - date when customer created an app account\n",
        "* gender (str) - gender of the customer (note some entries contain 'O' for other rather than M or F)\n",
        "* id (str) - customer id\n",
        "* income (float) - customer's income\n",
        "\n",
        "**transcript.json**\n",
        "* event (str) - record description (ie transaction, offer received, offer viewed, etc.)\n",
        "* person (str) - customer id\n",
        "* time (int) - time in hours since start of test. The data begins at time t=0\n",
        "* value - (dict of strings) - either an offer id or transaction amount depending on the record\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pU95MlWwyBHU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import math\n",
        "import json\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.preprocessing import PolynomialFeatures\n",
        "from sklearn.metrics import mean_squared_error\n",
        "from sklearn.metrics import classification_report\n",
        "from time import time\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.linear_model import Ridge\n",
        "from sklearn.tree import DecisionTreeRegressor\n",
        "\n",
        "% matplotlib inline\n",
        "\n",
        "# read in the json files\n",
        "portfolio = pd.read_json('data/portfolio.json', orient='records', lines=True)\n",
        "profile = pd.read_json('data/profile.json', orient='records', lines=True)\n",
        "transcript = pd.read_json('data/transcript.json', orient='records', lines=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OIrwrQ7EzMMS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "portfolio = pd.read_json('https://raw.githubusercontent.com/joshuayeung/Starbucks-Capstone-Challenge/master/data/portfolio.json', orient='records', lines=True)\n",
        "profile = pd.read_json('https://raw.githubusercontent.com/joshuayeung/Starbucks-Capstone-Challenge/master/data/profile.json', orient='records', lines=True)\n",
        "transcript = pd.read_json('https://raw.githubusercontent.com/joshuayeung/Starbucks-Capstone-Challenge/master/data/transcript.json', orient='records', lines=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xH5ChQC_yBHY",
        "colab_type": "text"
      },
      "source": [
        "# Data Exploration\n",
        "In order to analyze the problem better in next sections, first need to explore the datasets which includes checking the missing value, visualizing the data distribution, etc. In that way, we can have a better understanding on how the dataset looks like and how to select the important features to support the model implementation."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "84ulVfNXyBHY",
        "colab_type": "code",
        "colab": {},
        "outputId": "e3c2a82f-288a-4320-e59c-bcd120db3a7a"
      },
      "source": [
        "# view the portfolio dataset\n",
        "portfolio.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>duration</th>\n",
              "      <th>id</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>10</td>\n",
              "      <td>7</td>\n",
              "      <td>ae264e3637204a6fb9bb56bc8210ddfd</td>\n",
              "      <td>bogo</td>\n",
              "      <td>10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>10</td>\n",
              "      <td>5</td>\n",
              "      <td>4d5c57ea9a6940dd891ad53e9dbe8da0</td>\n",
              "      <td>bogo</td>\n",
              "      <td>10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>5</td>\n",
              "      <td>7</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>[web, email]</td>\n",
              "      <td>20</td>\n",
              "      <td>10</td>\n",
              "      <td>0b1e1539f2cc45b7b9fa7c272da2e1d7</td>\n",
              "      <td>discount</td>\n",
              "      <td>5</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                       channels  difficulty  duration  \\\n",
              "0       [email, mobile, social]          10         7   \n",
              "1  [web, email, mobile, social]          10         5   \n",
              "2          [web, email, mobile]           0         4   \n",
              "3          [web, email, mobile]           5         7   \n",
              "4                  [web, email]          20        10   \n",
              "\n",
              "                                 id     offer_type  reward  \n",
              "0  ae264e3637204a6fb9bb56bc8210ddfd           bogo      10  \n",
              "1  4d5c57ea9a6940dd891ad53e9dbe8da0           bogo      10  \n",
              "2  3f207df678b143eea3cee63160fa8bed  informational       0  \n",
              "3  9b98b8c7a33c4b65b9aebfe6a799e6d9           bogo       5  \n",
              "4  0b1e1539f2cc45b7b9fa7c272da2e1d7       discount       5  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E6tPLtS8yBHe",
        "colab_type": "code",
        "colab": {},
        "outputId": "950a66d5-59a2-499c-d2a6-f4befbf7fdb4"
      },
      "source": [
        "# quick check on missing value in dataset\n",
        "portfolio.isnull().sum()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "channels      0\n",
              "difficulty    0\n",
              "duration      0\n",
              "id            0\n",
              "offer_type    0\n",
              "reward        0\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fTrnyfWCyBHh",
        "colab_type": "text"
      },
      "source": [
        "As shown above, there's no missing values in portfolio dataset."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oetjw1JeyBHh",
        "colab_type": "code",
        "colab": {},
        "outputId": "54732c1c-e3de-4744-b435-d78b31dc596b"
      },
      "source": [
        "# view the shape of dateset\n",
        "portfolio.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(10, 6)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Los5HW0UyBHk",
        "colab_type": "code",
        "colab": {},
        "outputId": "cfe1703d-fdfd-43a5-b2fb-da7350406172"
      },
      "source": [
        "# the number of unique offer in dataset\n",
        "portfolio.id.nunique()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "10"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c8N3EazeyBHn",
        "colab_type": "code",
        "colab": {},
        "outputId": "695df627-ee30-4630-f211-b8e6dfbc554f"
      },
      "source": [
        "# count the number of offer by offer type\n",
        "portfolio.groupby('offer_type').id.count()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "offer_type\n",
              "bogo             4\n",
              "discount         4\n",
              "informational    2\n",
              "Name: id, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c5TXpu_eyBHq",
        "colab_type": "code",
        "colab": {},
        "outputId": "bce541b5-1d86-410f-9252-d4bc659d8938"
      },
      "source": [
        "# view the profile dataset which contains demographic information\n",
        "profile.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>age</th>\n",
              "      <th>became_member_on</th>\n",
              "      <th>gender</th>\n",
              "      <th>id</th>\n",
              "      <th>income</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>118</td>\n",
              "      <td>20170212</td>\n",
              "      <td>None</td>\n",
              "      <td>68be06ca386d4c31939f3a4f0e3dd783</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>55</td>\n",
              "      <td>20170715</td>\n",
              "      <td>F</td>\n",
              "      <td>0610b486422d4921ae7d2bf64640c50b</td>\n",
              "      <td>112000.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>118</td>\n",
              "      <td>20180712</td>\n",
              "      <td>None</td>\n",
              "      <td>38fe809add3b4fcf9315a9694bb96ff5</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>75</td>\n",
              "      <td>20170509</td>\n",
              "      <td>F</td>\n",
              "      <td>78afa995795e4d85b5d9ceeca43f5fef</td>\n",
              "      <td>100000.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>118</td>\n",
              "      <td>20170804</td>\n",
              "      <td>None</td>\n",
              "      <td>a03223e636434f42ac4c3df47e8bac43</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   age  became_member_on gender                                id    income\n",
              "0  118          20170212   None  68be06ca386d4c31939f3a4f0e3dd783       NaN\n",
              "1   55          20170715      F  0610b486422d4921ae7d2bf64640c50b  112000.0\n",
              "2  118          20180712   None  38fe809add3b4fcf9315a9694bb96ff5       NaN\n",
              "3   75          20170509      F  78afa995795e4d85b5d9ceeca43f5fef  100000.0\n",
              "4  118          20170804   None  a03223e636434f42ac4c3df47e8bac43       NaN"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "l6gB5f4-yBHy",
        "colab_type": "text"
      },
      "source": [
        "By viewing the first several rows of ther dataset, it parrently shows missing values in age column which is encoded as 118, and there are missing values in income column too."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fLM8PetVyBHz",
        "colab_type": "code",
        "colab": {},
        "outputId": "dd786a22-7144-44a7-d7e1-7f276e8cd1cc"
      },
      "source": [
        "# the number of user id\n",
        "profile['id'].nunique()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "17000"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P0WBFqEtyBH1",
        "colab_type": "code",
        "colab": {},
        "outputId": "d4b72722-db5f-4742-9c27-4f56dfb4db7e"
      },
      "source": [
        "# quick check for null values in dataset\n",
        "profile.isnull().sum()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "age                    0\n",
              "became_member_on       0\n",
              "gender              2175\n",
              "id                     0\n",
              "income              2175\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3MGfyuYkyBH4",
        "colab_type": "code",
        "colab": {},
        "outputId": "21c97476-cef5-4534-9d9d-2c49ff4e4b0e"
      },
      "source": [
        "# number of missing values in age column\n",
        "profile[profile.age==118].count()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "age                 2175\n",
              "became_member_on    2175\n",
              "gender                 0\n",
              "id                  2175\n",
              "income                 0\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-bh6-Yg4yBH6",
        "colab_type": "code",
        "colab": {},
        "outputId": "7440c7af-9dcc-43a9-c1f7-fe24ad08f859"
      },
      "source": [
        "# check if the rows which have missing age also have missing gender and income\n",
        "profile[profile.age==118].head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>age</th>\n",
              "      <th>became_member_on</th>\n",
              "      <th>gender</th>\n",
              "      <th>id</th>\n",
              "      <th>income</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>118</td>\n",
              "      <td>20170212</td>\n",
              "      <td>None</td>\n",
              "      <td>68be06ca386d4c31939f3a4f0e3dd783</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>118</td>\n",
              "      <td>20180712</td>\n",
              "      <td>None</td>\n",
              "      <td>38fe809add3b4fcf9315a9694bb96ff5</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>118</td>\n",
              "      <td>20170804</td>\n",
              "      <td>None</td>\n",
              "      <td>a03223e636434f42ac4c3df47e8bac43</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>118</td>\n",
              "      <td>20170925</td>\n",
              "      <td>None</td>\n",
              "      <td>8ec6ce2a7e7949b1bf142def7d0e0586</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>118</td>\n",
              "      <td>20171002</td>\n",
              "      <td>None</td>\n",
              "      <td>68617ca6246f4fbc85e91a2a49552598</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   age  became_member_on gender                                id  income\n",
              "0  118          20170212   None  68be06ca386d4c31939f3a4f0e3dd783     NaN\n",
              "2  118          20180712   None  38fe809add3b4fcf9315a9694bb96ff5     NaN\n",
              "4  118          20170804   None  a03223e636434f42ac4c3df47e8bac43     NaN\n",
              "6  118          20170925   None  8ec6ce2a7e7949b1bf142def7d0e0586     NaN\n",
              "7  118          20171002   None  68617ca6246f4fbc85e91a2a49552598     NaN"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p97xLlTayBH9",
        "colab_type": "text"
      },
      "source": [
        "Apparently, the rows which have missing age also missing gender and income, which means probably it's fine to just drop the rows in the following steps to support the model implementation."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VxvDJkI1yBH9",
        "colab_type": "code",
        "colab": {},
        "outputId": "538bded8-3a6f-4276-ff7c-778ccb81b942"
      },
      "source": [
        "# get a quick view on income distribution\n",
        "profile.income.hist()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f2d1b05e160>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f2d1b47e2e8>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uKX2avAFyBIB",
        "colab_type": "code",
        "colab": {},
        "outputId": "814ff34c-4368-4ca3-aeab-c94a61ff4498"
      },
      "source": [
        "# view the transcript dataset\n",
        "transcript.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>offer received</td>\n",
              "      <td>78afa995795e4d85b5d9ceeca43f5fef</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '9b98b8c7a33c4b65b9aebfe6a799e6d9'}</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer received</td>\n",
              "      <td>a03223e636434f42ac4c3df47e8bac43</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '0b1e1539f2cc45b7b9fa7c272da2e1d7'}</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>offer received</td>\n",
              "      <td>e2127556f4f64592b11af22de27a7932</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '2906b810c7d4411798c6938adc9daaa5'}</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>offer received</td>\n",
              "      <td>8ec6ce2a7e7949b1bf142def7d0e0586</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer received</td>\n",
              "      <td>68617ca6246f4fbc85e91a2a49552598</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '4d5c57ea9a6940dd891ad53e9dbe8da0'}</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            event                            person  time  \\\n",
              "0  offer received  78afa995795e4d85b5d9ceeca43f5fef     0   \n",
              "1  offer received  a03223e636434f42ac4c3df47e8bac43     0   \n",
              "2  offer received  e2127556f4f64592b11af22de27a7932     0   \n",
              "3  offer received  8ec6ce2a7e7949b1bf142def7d0e0586     0   \n",
              "4  offer received  68617ca6246f4fbc85e91a2a49552598     0   \n",
              "\n",
              "                                              value  \n",
              "0  {'offer id': '9b98b8c7a33c4b65b9aebfe6a799e6d9'}  \n",
              "1  {'offer id': '0b1e1539f2cc45b7b9fa7c272da2e1d7'}  \n",
              "2  {'offer id': '2906b810c7d4411798c6938adc9daaa5'}  \n",
              "3  {'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}  \n",
              "4  {'offer id': '4d5c57ea9a6940dd891ad53e9dbe8da0'}  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jZi2ecIpyBIE",
        "colab_type": "code",
        "colab": {},
        "outputId": "3d287351-b548-48d1-e9f4-1a3f8a6b51a2"
      },
      "source": [
        "# quick check on missing value\n",
        "transcript.isnull().sum()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "event     0\n",
              "person    0\n",
              "time      0\n",
              "value     0\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XviT-HhvyBIG",
        "colab_type": "code",
        "colab": {},
        "outputId": "b6dea53c-537c-4ba2-8556-96cc80ed4e06"
      },
      "source": [
        "# quick chekc on the event distribution\n",
        "transcript.groupby('event').person.count()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "event\n",
              "offer completed     33579\n",
              "offer received      76277\n",
              "offer viewed        57725\n",
              "transaction        138953\n",
              "Name: person, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CcvqonFTyBIJ",
        "colab_type": "code",
        "colab": {},
        "outputId": "b5218d30-6158-48e7-def0-e76a29ce3494"
      },
      "source": [
        "# extract the different values in value column out\n",
        "transcript = pd.concat([transcript, transcript['value'].apply(pd.Series)], axis=1)\n",
        "transcript.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>offer id</th>\n",
              "      <th>amount</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>reward</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>offer received</td>\n",
              "      <td>78afa995795e4d85b5d9ceeca43f5fef</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '9b98b8c7a33c4b65b9aebfe6a799e6d9'}</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer received</td>\n",
              "      <td>a03223e636434f42ac4c3df47e8bac43</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '0b1e1539f2cc45b7b9fa7c272da2e1d7'}</td>\n",
              "      <td>0b1e1539f2cc45b7b9fa7c272da2e1d7</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>offer received</td>\n",
              "      <td>e2127556f4f64592b11af22de27a7932</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '2906b810c7d4411798c6938adc9daaa5'}</td>\n",
              "      <td>2906b810c7d4411798c6938adc9daaa5</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>offer received</td>\n",
              "      <td>8ec6ce2a7e7949b1bf142def7d0e0586</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer received</td>\n",
              "      <td>68617ca6246f4fbc85e91a2a49552598</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '4d5c57ea9a6940dd891ad53e9dbe8da0'}</td>\n",
              "      <td>4d5c57ea9a6940dd891ad53e9dbe8da0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            event                            person  time  \\\n",
              "0  offer received  78afa995795e4d85b5d9ceeca43f5fef     0   \n",
              "1  offer received  a03223e636434f42ac4c3df47e8bac43     0   \n",
              "2  offer received  e2127556f4f64592b11af22de27a7932     0   \n",
              "3  offer received  8ec6ce2a7e7949b1bf142def7d0e0586     0   \n",
              "4  offer received  68617ca6246f4fbc85e91a2a49552598     0   \n",
              "\n",
              "                                              value  \\\n",
              "0  {'offer id': '9b98b8c7a33c4b65b9aebfe6a799e6d9'}   \n",
              "1  {'offer id': '0b1e1539f2cc45b7b9fa7c272da2e1d7'}   \n",
              "2  {'offer id': '2906b810c7d4411798c6938adc9daaa5'}   \n",
              "3  {'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}   \n",
              "4  {'offer id': '4d5c57ea9a6940dd891ad53e9dbe8da0'}   \n",
              "\n",
              "                           offer id  amount offer_id  reward  \n",
              "0  9b98b8c7a33c4b65b9aebfe6a799e6d9     NaN      NaN     NaN  \n",
              "1  0b1e1539f2cc45b7b9fa7c272da2e1d7     NaN      NaN     NaN  \n",
              "2  2906b810c7d4411798c6938adc9daaa5     NaN      NaN     NaN  \n",
              "3  fafdcd668e3743c1bb461111dcafc2a4     NaN      NaN     NaN  \n",
              "4  4d5c57ea9a6940dd891ad53e9dbe8da0     NaN      NaN     NaN  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HyeucuKkyBIM",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# clean up the duplicates in offer id and offer_id and meger into one column\n",
        "transcript['clean_id'] = np.where(transcript['offer id'].isnull() & transcript['offer_id'].notnull(), transcript['offer_id'],transcript['offer id'])\n",
        "\n",
        "# drop the original id columns\n",
        "transcript.drop(['offer id', 'offer_id'], axis=1, inplace=True)\n",
        "\n",
        "# rename the offer_id column\n",
        "transcript.rename(columns={'clean_id': 'offer_id'}, inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fKAh0RJ2yBIQ",
        "colab_type": "code",
        "colab": {},
        "outputId": "2c3879f0-f86d-4b50-f637-0855c1b9a46f"
      },
      "source": [
        "# quick check on cleaned dataframe\n",
        "transcript.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>reward</th>\n",
              "      <th>offer_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>offer received</td>\n",
              "      <td>78afa995795e4d85b5d9ceeca43f5fef</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '9b98b8c7a33c4b65b9aebfe6a799e6d9'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer received</td>\n",
              "      <td>a03223e636434f42ac4c3df47e8bac43</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '0b1e1539f2cc45b7b9fa7c272da2e1d7'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0b1e1539f2cc45b7b9fa7c272da2e1d7</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>offer received</td>\n",
              "      <td>e2127556f4f64592b11af22de27a7932</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '2906b810c7d4411798c6938adc9daaa5'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2906b810c7d4411798c6938adc9daaa5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>offer received</td>\n",
              "      <td>8ec6ce2a7e7949b1bf142def7d0e0586</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer received</td>\n",
              "      <td>68617ca6246f4fbc85e91a2a49552598</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '4d5c57ea9a6940dd891ad53e9dbe8da0'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>4d5c57ea9a6940dd891ad53e9dbe8da0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            event                            person  time  \\\n",
              "0  offer received  78afa995795e4d85b5d9ceeca43f5fef     0   \n",
              "1  offer received  a03223e636434f42ac4c3df47e8bac43     0   \n",
              "2  offer received  e2127556f4f64592b11af22de27a7932     0   \n",
              "3  offer received  8ec6ce2a7e7949b1bf142def7d0e0586     0   \n",
              "4  offer received  68617ca6246f4fbc85e91a2a49552598     0   \n",
              "\n",
              "                                              value  amount  reward  \\\n",
              "0  {'offer id': '9b98b8c7a33c4b65b9aebfe6a799e6d9'}     NaN     NaN   \n",
              "1  {'offer id': '0b1e1539f2cc45b7b9fa7c272da2e1d7'}     NaN     NaN   \n",
              "2  {'offer id': '2906b810c7d4411798c6938adc9daaa5'}     NaN     NaN   \n",
              "3  {'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}     NaN     NaN   \n",
              "4  {'offer id': '4d5c57ea9a6940dd891ad53e9dbe8da0'}     NaN     NaN   \n",
              "\n",
              "                           offer_id  \n",
              "0  9b98b8c7a33c4b65b9aebfe6a799e6d9  \n",
              "1  0b1e1539f2cc45b7b9fa7c272da2e1d7  \n",
              "2  2906b810c7d4411798c6938adc9daaa5  \n",
              "3  fafdcd668e3743c1bb461111dcafc2a4  \n",
              "4  4d5c57ea9a6940dd891ad53e9dbe8da0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HTvSivVAyBIU",
        "colab_type": "code",
        "colab": {},
        "outputId": "ce72ab11-4b74-47fb-a130-ee6e10b43c57"
      },
      "source": [
        "# check if the number of type of offer consist with the portfolio dataset\n",
        "transcript.offer_id.nunique()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "10"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dZ2XwUubyBIW",
        "colab_type": "text"
      },
      "source": [
        "# Data Preprocessing\n",
        "In order to find out what mainly affect the finish of the transaction by sending the offer, in the data processing process, also need to process the data to merge the events of each specific offer sent so as to find out which offer were received, viewed and finally compeleted with a transaction."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": true,
        "id": "kEHK_Dg3yBIX",
        "colab_type": "code",
        "colab": {},
        "outputId": "4710dacb-ac87-457c-b7c2-33250c1ef84c"
      },
      "source": [
        "# quick view on event distribution by offer_type by combining two dataset (portfolio, transcript)\n",
        "transcript=transcript.merge(portfolio,how='left',left_on='offer_id',right_on='id')\n",
        "transcript.groupby(['event','offer_type'])['offer_type'].count()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "event            offer_type   \n",
              "offer completed  bogo             15669\n",
              "                 discount         17910\n",
              "offer received   bogo             30499\n",
              "                 discount         30543\n",
              "                 informational    15235\n",
              "offer viewed     bogo             25449\n",
              "                 discount         21445\n",
              "                 informational    10831\n",
              "Name: offer_type, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M-9sY52ryBIZ",
        "colab_type": "text"
      },
      "source": [
        "Since offer_id is not associated with any 'transaction' event, in order to flag whether the offer has been finally completed with a transaction, here we need to link the offer id back to all transaction events. For BOGO and discount offer, both of them will have the consequence of offer received, viewed, transaction and offer completed which will apparently show that the offer is redeemed and should definitely be sent out. For the information offer, though there's no reward step there should still be a transaction which is linked to the usage of the offer."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UVxE0UjNyBIZ",
        "colab_type": "code",
        "colab": {},
        "outputId": "ec566abf-5af4-4355-de77-64cafc089aa1"
      },
      "source": [
        "# flag the customer who is really responsive to the offer\n",
        "# (only when customer do the transaction after receiving and viewing the offer, it chould be counted as being \n",
        "# responsive to the offer, otherwise either 1) viewing the offer without any transaction or 2) make the transaction\n",
        "# without receiving or viewing the offer shouldn't be counted as a response)\n",
        "\n",
        "# drop the useless columns\n",
        "transcript_processed = pd.DataFrame(transcript)\n",
        "\n",
        "transcript_processed = transcript_processed.drop(columns=['reward_x','reward_y'], axis=1,inplace=False)\n",
        "transcript_processed.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>duration</th>\n",
              "      <th>id</th>\n",
              "      <th>offer_type</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>offer received</td>\n",
              "      <td>78afa995795e4d85b5d9ceeca43f5fef</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '9b98b8c7a33c4b65b9aebfe6a799e6d9'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>5.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>bogo</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer received</td>\n",
              "      <td>a03223e636434f42ac4c3df47e8bac43</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '0b1e1539f2cc45b7b9fa7c272da2e1d7'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0b1e1539f2cc45b7b9fa7c272da2e1d7</td>\n",
              "      <td>[web, email]</td>\n",
              "      <td>20.0</td>\n",
              "      <td>10.0</td>\n",
              "      <td>0b1e1539f2cc45b7b9fa7c272da2e1d7</td>\n",
              "      <td>discount</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>offer received</td>\n",
              "      <td>e2127556f4f64592b11af22de27a7932</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '2906b810c7d4411798c6938adc9daaa5'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2906b810c7d4411798c6938adc9daaa5</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>10.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>2906b810c7d4411798c6938adc9daaa5</td>\n",
              "      <td>discount</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>offer received</td>\n",
              "      <td>8ec6ce2a7e7949b1bf142def7d0e0586</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>10.0</td>\n",
              "      <td>10.0</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>discount</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer received</td>\n",
              "      <td>68617ca6246f4fbc85e91a2a49552598</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '4d5c57ea9a6940dd891ad53e9dbe8da0'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>4d5c57ea9a6940dd891ad53e9dbe8da0</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>10.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>4d5c57ea9a6940dd891ad53e9dbe8da0</td>\n",
              "      <td>bogo</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            event                            person  time  \\\n",
              "0  offer received  78afa995795e4d85b5d9ceeca43f5fef     0   \n",
              "1  offer received  a03223e636434f42ac4c3df47e8bac43     0   \n",
              "2  offer received  e2127556f4f64592b11af22de27a7932     0   \n",
              "3  offer received  8ec6ce2a7e7949b1bf142def7d0e0586     0   \n",
              "4  offer received  68617ca6246f4fbc85e91a2a49552598     0   \n",
              "\n",
              "                                              value  amount  \\\n",
              "0  {'offer id': '9b98b8c7a33c4b65b9aebfe6a799e6d9'}     NaN   \n",
              "1  {'offer id': '0b1e1539f2cc45b7b9fa7c272da2e1d7'}     NaN   \n",
              "2  {'offer id': '2906b810c7d4411798c6938adc9daaa5'}     NaN   \n",
              "3  {'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}     NaN   \n",
              "4  {'offer id': '4d5c57ea9a6940dd891ad53e9dbe8da0'}     NaN   \n",
              "\n",
              "                           offer_id                      channels  difficulty  \\\n",
              "0  9b98b8c7a33c4b65b9aebfe6a799e6d9          [web, email, mobile]         5.0   \n",
              "1  0b1e1539f2cc45b7b9fa7c272da2e1d7                  [web, email]        20.0   \n",
              "2  2906b810c7d4411798c6938adc9daaa5          [web, email, mobile]        10.0   \n",
              "3  fafdcd668e3743c1bb461111dcafc2a4  [web, email, mobile, social]        10.0   \n",
              "4  4d5c57ea9a6940dd891ad53e9dbe8da0  [web, email, mobile, social]        10.0   \n",
              "\n",
              "   duration                                id offer_type  \n",
              "0       7.0  9b98b8c7a33c4b65b9aebfe6a799e6d9       bogo  \n",
              "1      10.0  0b1e1539f2cc45b7b9fa7c272da2e1d7   discount  \n",
              "2       7.0  2906b810c7d4411798c6938adc9daaa5   discount  \n",
              "3      10.0  fafdcd668e3743c1bb461111dcafc2a4   discount  \n",
              "4       5.0  4d5c57ea9a6940dd891ad53e9dbe8da0       bogo  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": false,
        "id": "LMhBspN7yBIc",
        "colab_type": "code",
        "colab": {},
        "outputId": "768b41c9-25c1-467a-b933-184b95e05203"
      },
      "source": [
        "# sort the events by person and time\n",
        "transcript_processed = transcript_processed.sort_values(['person','time'])\n",
        "transcript_processed.head(10)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>duration</th>\n",
              "      <th>id</th>\n",
              "      <th>offer_type</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>55972</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>168</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>77705</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>192</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>89291</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>228</td>\n",
              "      <td>{'amount': 22.16}</td>\n",
              "      <td>22.16</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>113605</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>336</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>139992</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>372</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>153401</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>408</td>\n",
              "      <td>{'offer id': 'f19421c1d4aa40978ebb69ca19b0e20d'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>5.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>bogo</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>168412</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>414</td>\n",
              "      <td>{'amount': 8.57}</td>\n",
              "      <td>8.57</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>168413</th>\n",
              "      <td>offer completed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>414</td>\n",
              "      <td>{'offer_id': 'f19421c1d4aa40978ebb69ca19b0e20d...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>5.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>bogo</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>187554</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>456</td>\n",
              "      <td>{'offer id': 'f19421c1d4aa40978ebb69ca19b0e20d'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>5.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>bogo</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>204340</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>504</td>\n",
              "      <td>{'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>10.0</td>\n",
              "      <td>10.0</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>discount</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                  event                            person  time  \\\n",
              "55972    offer received  0009655768c64bdeb2e877511632db8f   168   \n",
              "77705      offer viewed  0009655768c64bdeb2e877511632db8f   192   \n",
              "89291       transaction  0009655768c64bdeb2e877511632db8f   228   \n",
              "113605   offer received  0009655768c64bdeb2e877511632db8f   336   \n",
              "139992     offer viewed  0009655768c64bdeb2e877511632db8f   372   \n",
              "153401   offer received  0009655768c64bdeb2e877511632db8f   408   \n",
              "168412      transaction  0009655768c64bdeb2e877511632db8f   414   \n",
              "168413  offer completed  0009655768c64bdeb2e877511632db8f   414   \n",
              "187554     offer viewed  0009655768c64bdeb2e877511632db8f   456   \n",
              "204340   offer received  0009655768c64bdeb2e877511632db8f   504   \n",
              "\n",
              "                                                    value  amount  \\\n",
              "55972    {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "77705    {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "89291                                   {'amount': 22.16}   22.16   \n",
              "113605   {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "139992   {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "153401   {'offer id': 'f19421c1d4aa40978ebb69ca19b0e20d'}     NaN   \n",
              "168412                                   {'amount': 8.57}    8.57   \n",
              "168413  {'offer_id': 'f19421c1d4aa40978ebb69ca19b0e20d...     NaN   \n",
              "187554   {'offer id': 'f19421c1d4aa40978ebb69ca19b0e20d'}     NaN   \n",
              "204340   {'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}     NaN   \n",
              "\n",
              "                                offer_id                      channels  \\\n",
              "55972   5a8bc65990b245e5a138643cd4eb9837       [email, mobile, social]   \n",
              "77705   5a8bc65990b245e5a138643cd4eb9837       [email, mobile, social]   \n",
              "89291                                NaN                           NaN   \n",
              "113605  3f207df678b143eea3cee63160fa8bed          [web, email, mobile]   \n",
              "139992  3f207df678b143eea3cee63160fa8bed          [web, email, mobile]   \n",
              "153401  f19421c1d4aa40978ebb69ca19b0e20d  [web, email, mobile, social]   \n",
              "168412                               NaN                           NaN   \n",
              "168413  f19421c1d4aa40978ebb69ca19b0e20d  [web, email, mobile, social]   \n",
              "187554  f19421c1d4aa40978ebb69ca19b0e20d  [web, email, mobile, social]   \n",
              "204340  fafdcd668e3743c1bb461111dcafc2a4  [web, email, mobile, social]   \n",
              "\n",
              "        difficulty  duration                                id     offer_type  \n",
              "55972          0.0       3.0  5a8bc65990b245e5a138643cd4eb9837  informational  \n",
              "77705          0.0       3.0  5a8bc65990b245e5a138643cd4eb9837  informational  \n",
              "89291          NaN       NaN                               NaN            NaN  \n",
              "113605         0.0       4.0  3f207df678b143eea3cee63160fa8bed  informational  \n",
              "139992         0.0       4.0  3f207df678b143eea3cee63160fa8bed  informational  \n",
              "153401         5.0       5.0  f19421c1d4aa40978ebb69ca19b0e20d           bogo  \n",
              "168412         NaN       NaN                               NaN            NaN  \n",
              "168413         5.0       5.0  f19421c1d4aa40978ebb69ca19b0e20d           bogo  \n",
              "187554         5.0       5.0  f19421c1d4aa40978ebb69ca19b0e20d           bogo  \n",
              "204340        10.0      10.0  fafdcd668e3743c1bb461111dcafc2a4       discount  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EIBSg49EyBIi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "offer_completed = transcript_processed[['time','offer_id','person','event']][(transcript_processed['event']\\\n",
        "    =='transaction') | (transcript_processed['event']=='offer viewed')].groupby(['person','offer_id']).ffill()\n",
        "offer_completed['offer_id'] = offer_completed['offer_id'].ffill()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PiTAfYdlyBIm",
        "colab_type": "code",
        "colab": {},
        "outputId": "446fa38d-2481-462a-d380-b49b91abb16f"
      },
      "source": [
        "# create the complete whole dataset for filtered data\n",
        "transcript_processed = transcript_processed.merge(offer_completed, how='left', on=['person','time','event'])\n",
        "transcript_processed.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>offer_id_x</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>duration</th>\n",
              "      <th>id</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>offer_id_y</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>168</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>192</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>228</td>\n",
              "      <td>{'amount': 22.16}</td>\n",
              "      <td>22.16</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>336</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>372</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            event                            person  time  \\\n",
              "0  offer received  0009655768c64bdeb2e877511632db8f   168   \n",
              "1    offer viewed  0009655768c64bdeb2e877511632db8f   192   \n",
              "2     transaction  0009655768c64bdeb2e877511632db8f   228   \n",
              "3  offer received  0009655768c64bdeb2e877511632db8f   336   \n",
              "4    offer viewed  0009655768c64bdeb2e877511632db8f   372   \n",
              "\n",
              "                                              value  amount  \\\n",
              "0  {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "1  {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "2                                 {'amount': 22.16}   22.16   \n",
              "3  {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "4  {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "\n",
              "                         offer_id_x                 channels  difficulty  \\\n",
              "0  5a8bc65990b245e5a138643cd4eb9837  [email, mobile, social]         0.0   \n",
              "1  5a8bc65990b245e5a138643cd4eb9837  [email, mobile, social]         0.0   \n",
              "2                               NaN                      NaN         NaN   \n",
              "3  3f207df678b143eea3cee63160fa8bed     [web, email, mobile]         0.0   \n",
              "4  3f207df678b143eea3cee63160fa8bed     [web, email, mobile]         0.0   \n",
              "\n",
              "   duration                                id     offer_type  \\\n",
              "0       3.0  5a8bc65990b245e5a138643cd4eb9837  informational   \n",
              "1       3.0  5a8bc65990b245e5a138643cd4eb9837  informational   \n",
              "2       NaN                               NaN            NaN   \n",
              "3       4.0  3f207df678b143eea3cee63160fa8bed  informational   \n",
              "4       4.0  3f207df678b143eea3cee63160fa8bed  informational   \n",
              "\n",
              "                         offer_id_y  \n",
              "0                               NaN  \n",
              "1  5a8bc65990b245e5a138643cd4eb9837  \n",
              "2  5a8bc65990b245e5a138643cd4eb9837  \n",
              "3                               NaN  \n",
              "4  3f207df678b143eea3cee63160fa8bed  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HKOv1GH_yBIo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# clean up the dataframe (removing the dummy columns)\n",
        "transcript_processed['offer_id'] = np.where(transcript_processed['offer_id_x'].isnull(),\\\n",
        "                                            transcript_processed['offer_id_y'], transcript_processed['offer_id_x'])\n",
        "transcript_processed.drop(columns=['offer_id_x','offer_id_y'], axis=1, inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gP49La-nyBIq",
        "colab_type": "code",
        "colab": {},
        "outputId": "ce0bd250-41ed-454f-f9e5-27688d12cebc"
      },
      "source": [
        "transcript_processed.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>duration</th>\n",
              "      <th>id</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>offer_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>168</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>192</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>228</td>\n",
              "      <td>{'amount': 22.16}</td>\n",
              "      <td>22.16</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>336</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>372</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            event                            person  time  \\\n",
              "0  offer received  0009655768c64bdeb2e877511632db8f   168   \n",
              "1    offer viewed  0009655768c64bdeb2e877511632db8f   192   \n",
              "2     transaction  0009655768c64bdeb2e877511632db8f   228   \n",
              "3  offer received  0009655768c64bdeb2e877511632db8f   336   \n",
              "4    offer viewed  0009655768c64bdeb2e877511632db8f   372   \n",
              "\n",
              "                                              value  amount  \\\n",
              "0  {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "1  {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "2                                 {'amount': 22.16}   22.16   \n",
              "3  {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "4  {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "\n",
              "                  channels  difficulty  duration  \\\n",
              "0  [email, mobile, social]         0.0       3.0   \n",
              "1  [email, mobile, social]         0.0       3.0   \n",
              "2                      NaN         NaN       NaN   \n",
              "3     [web, email, mobile]         0.0       4.0   \n",
              "4     [web, email, mobile]         0.0       4.0   \n",
              "\n",
              "                                 id     offer_type  \\\n",
              "0  5a8bc65990b245e5a138643cd4eb9837  informational   \n",
              "1  5a8bc65990b245e5a138643cd4eb9837  informational   \n",
              "2                               NaN            NaN   \n",
              "3  3f207df678b143eea3cee63160fa8bed  informational   \n",
              "4  3f207df678b143eea3cee63160fa8bed  informational   \n",
              "\n",
              "                           offer_id  \n",
              "0  5a8bc65990b245e5a138643cd4eb9837  \n",
              "1  5a8bc65990b245e5a138643cd4eb9837  \n",
              "2  5a8bc65990b245e5a138643cd4eb9837  \n",
              "3  3f207df678b143eea3cee63160fa8bed  \n",
              "4  3f207df678b143eea3cee63160fa8bed  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Yv1yAYLKyBIt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "transcript_processed = transcript_processed.merge(portfolio, how = 'left', left_on='offer_id', right_on='id')\n",
        "transcript_processed['duration'] = np.where(transcript_processed['duration_x'].isnull(), \\\n",
        "                                            transcript_processed['duration_y'], transcript_processed['duration_x'])\n",
        "transcript_processed.drop(columns=['duration_x','offer_type_x','difficulty_x','channels_x','duration_y'],\\\n",
        "                         axis=1, inplace=True)\n",
        "transcript_processed.rename(columns={'channels_y':'channels','reward_y':'reward','difficulty_y':'difficulty','offer_type_y':'offer_type'},inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": true,
        "id": "kx4l0hkTyBIv",
        "colab_type": "code",
        "colab": {},
        "outputId": "f54f532f-5abc-4ef0-ad49-8904751153bf"
      },
      "source": [
        "# quick check on processed dataset\n",
        "transcript_processed.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>id_x</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>id_y</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward</th>\n",
              "      <th>duration</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>168</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>192</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>228</td>\n",
              "      <td>{'amount': 22.16}</td>\n",
              "      <td>22.16</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>336</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>4.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>372</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>4.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            event                            person  time  \\\n",
              "0  offer received  0009655768c64bdeb2e877511632db8f   168   \n",
              "1    offer viewed  0009655768c64bdeb2e877511632db8f   192   \n",
              "2     transaction  0009655768c64bdeb2e877511632db8f   228   \n",
              "3  offer received  0009655768c64bdeb2e877511632db8f   336   \n",
              "4    offer viewed  0009655768c64bdeb2e877511632db8f   372   \n",
              "\n",
              "                                              value  amount  \\\n",
              "0  {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "1  {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "2                                 {'amount': 22.16}   22.16   \n",
              "3  {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "4  {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "\n",
              "                               id_x                          offer_id  \\\n",
              "0  5a8bc65990b245e5a138643cd4eb9837  5a8bc65990b245e5a138643cd4eb9837   \n",
              "1  5a8bc65990b245e5a138643cd4eb9837  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2                               NaN  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3  3f207df678b143eea3cee63160fa8bed  3f207df678b143eea3cee63160fa8bed   \n",
              "4  3f207df678b143eea3cee63160fa8bed  3f207df678b143eea3cee63160fa8bed   \n",
              "\n",
              "                  channels  difficulty                              id_y  \\\n",
              "0  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "1  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3     [web, email, mobile]           0  3f207df678b143eea3cee63160fa8bed   \n",
              "4     [web, email, mobile]           0  3f207df678b143eea3cee63160fa8bed   \n",
              "\n",
              "      offer_type  reward  duration  \n",
              "0  informational       0       3.0  \n",
              "1  informational       0       3.0  \n",
              "2  informational       0       3.0  \n",
              "3  informational       0       4.0  \n",
              "4  informational       0       4.0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6H_zdcT3yBI0",
        "colab_type": "text"
      },
      "source": [
        "Next, after we get the data together, we need to extract the transactions which were completed after the offer was received and viewed. Since we've already filled all transaction's offer id, we can extract the transactions converted from offers by checking if the offer id before the transaction is the same as the transaction's offer id."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JLqgsG0qyBI0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# subset the dataset with only offer viewed, transaction, and offer completed events\n",
        "transactions_after_viewed = transcript_processed[(transcript_processed['event']=='offer viewed')|\\\n",
        "                                                 (transcript_processed['event']=='transaction')|\\\n",
        "                                                 (transcript_processed['event']=='offer completed')].copy()\n",
        "\n",
        "# generate the previous offer id\n",
        "transactions_after_viewed['pre_offer_id'] = transactions_after_viewed.groupby(['person', 'offer_id'])['offer_id'].shift()\n",
        "\n",
        "# create flag for responsed offer which competed after customer viewing the offer\n",
        "transactions_after_viewed['completed_offer'] = np.where(transactions_after_viewed['pre_offer_id']==\\\n",
        "                                                       transactions_after_viewed['offer_id'],1,0)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "REb9s0lbyBI2",
        "colab_type": "code",
        "colab": {},
        "outputId": "dbbbb1db-5e3c-45a7-d304-2253d7f2cb92"
      },
      "source": [
        "# quick view on dataset\n",
        "transactions_after_viewed.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>id_x</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>id_y</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward</th>\n",
              "      <th>duration</th>\n",
              "      <th>pre_offer_id</th>\n",
              "      <th>completed_offer</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>192</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>228</td>\n",
              "      <td>{'amount': 22.16}</td>\n",
              "      <td>22.16</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>372</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>414</td>\n",
              "      <td>{'amount': 8.57}</td>\n",
              "      <td>8.57</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>offer completed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>414</td>\n",
              "      <td>{'offer_id': 'f19421c1d4aa40978ebb69ca19b0e20d...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>5</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5</td>\n",
              "      <td>5.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "             event                            person  time  \\\n",
              "1     offer viewed  0009655768c64bdeb2e877511632db8f   192   \n",
              "2      transaction  0009655768c64bdeb2e877511632db8f   228   \n",
              "4     offer viewed  0009655768c64bdeb2e877511632db8f   372   \n",
              "6      transaction  0009655768c64bdeb2e877511632db8f   414   \n",
              "7  offer completed  0009655768c64bdeb2e877511632db8f   414   \n",
              "\n",
              "                                               value  amount  \\\n",
              "1   {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "2                                  {'amount': 22.16}   22.16   \n",
              "4   {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "6                                   {'amount': 8.57}    8.57   \n",
              "7  {'offer_id': 'f19421c1d4aa40978ebb69ca19b0e20d...     NaN   \n",
              "\n",
              "                               id_x                          offer_id  \\\n",
              "1  5a8bc65990b245e5a138643cd4eb9837  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2                               NaN  5a8bc65990b245e5a138643cd4eb9837   \n",
              "4  3f207df678b143eea3cee63160fa8bed  3f207df678b143eea3cee63160fa8bed   \n",
              "6                               NaN  3f207df678b143eea3cee63160fa8bed   \n",
              "7  f19421c1d4aa40978ebb69ca19b0e20d  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "\n",
              "                       channels  difficulty                              id_y  \\\n",
              "1       [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2       [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "4          [web, email, mobile]           0  3f207df678b143eea3cee63160fa8bed   \n",
              "6          [web, email, mobile]           0  3f207df678b143eea3cee63160fa8bed   \n",
              "7  [web, email, mobile, social]           5  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "\n",
              "      offer_type  reward  duration                      pre_offer_id  \\\n",
              "1  informational       0       3.0                               NaN   \n",
              "2  informational       0       3.0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "4  informational       0       4.0                               NaN   \n",
              "6  informational       0       4.0  3f207df678b143eea3cee63160fa8bed   \n",
              "7           bogo       5       5.0                               NaN   \n",
              "\n",
              "   completed_offer  \n",
              "1                0  \n",
              "2                1  \n",
              "4                0  \n",
              "6                1  \n",
              "7                0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yYSXqkMQyBI5",
        "colab_type": "code",
        "colab": {},
        "outputId": "90d4a4ed-f953-4d80-aee4-217de2f84e2d"
      },
      "source": [
        "# join back the 'offer received' events which was filtered out in the previous step\n",
        "offer_received = transcript_processed[transcript_processed['event']=='offer received']\n",
        "\n",
        "offer_received['pre_offer_id']=np.nan\n",
        "offer_received['completed_offer']=np.nan\n",
        "\n",
        "transcript_processed = offer_received.append(transactions_after_viewed).sort_values(['person','time'])\n",
        "transcript_processed.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/opt/conda/lib/python3.6/site-packages/ipykernel_launcher.py:4: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame.\n",
            "Try using .loc[row_indexer,col_indexer] = value instead\n",
            "\n",
            "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
            "  after removing the cwd from sys.path.\n",
            "/opt/conda/lib/python3.6/site-packages/ipykernel_launcher.py:5: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame.\n",
            "Try using .loc[row_indexer,col_indexer] = value instead\n",
            "\n",
            "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
            "  \"\"\"\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>id_x</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>id_y</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward</th>\n",
              "      <th>duration</th>\n",
              "      <th>pre_offer_id</th>\n",
              "      <th>completed_offer</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>168</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>192</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>228</td>\n",
              "      <td>{'amount': 22.16}</td>\n",
              "      <td>22.16</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>336</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>372</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            event                            person  time  \\\n",
              "0  offer received  0009655768c64bdeb2e877511632db8f   168   \n",
              "1    offer viewed  0009655768c64bdeb2e877511632db8f   192   \n",
              "2     transaction  0009655768c64bdeb2e877511632db8f   228   \n",
              "3  offer received  0009655768c64bdeb2e877511632db8f   336   \n",
              "4    offer viewed  0009655768c64bdeb2e877511632db8f   372   \n",
              "\n",
              "                                              value  amount  \\\n",
              "0  {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "1  {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "2                                 {'amount': 22.16}   22.16   \n",
              "3  {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "4  {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "\n",
              "                               id_x                          offer_id  \\\n",
              "0  5a8bc65990b245e5a138643cd4eb9837  5a8bc65990b245e5a138643cd4eb9837   \n",
              "1  5a8bc65990b245e5a138643cd4eb9837  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2                               NaN  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3  3f207df678b143eea3cee63160fa8bed  3f207df678b143eea3cee63160fa8bed   \n",
              "4  3f207df678b143eea3cee63160fa8bed  3f207df678b143eea3cee63160fa8bed   \n",
              "\n",
              "                  channels  difficulty                              id_y  \\\n",
              "0  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "1  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3     [web, email, mobile]           0  3f207df678b143eea3cee63160fa8bed   \n",
              "4     [web, email, mobile]           0  3f207df678b143eea3cee63160fa8bed   \n",
              "\n",
              "      offer_type  reward  duration                      pre_offer_id  \\\n",
              "0  informational       0       3.0                               NaN   \n",
              "1  informational       0       3.0                               NaN   \n",
              "2  informational       0       3.0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3  informational       0       4.0                               NaN   \n",
              "4  informational       0       4.0                               NaN   \n",
              "\n",
              "   completed_offer  \n",
              "0              NaN  \n",
              "1              0.0  \n",
              "2              1.0  \n",
              "3              NaN  \n",
              "4              0.0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hTPdmFrtyBI7",
        "colab_type": "text"
      },
      "source": [
        "Since the different offer has difference consequence of completion, for example, for the informational offer, there'll not be rewards. Therefore, separate the transcript data by offer type for easier analysis."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OsoU-fdnyBI7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#split transcript into 3 different offer types\n",
        "bogo = transcript_processed[transcript_processed['offer_type']=='bogo'].copy()\n",
        "discount = transcript_processed[transcript_processed['offer_type']=='discount'].copy()\n",
        "informational = transcript_processed[transcript_processed['offer_type']=='informational'].copy()\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5-qiOfA_yBI9",
        "colab_type": "code",
        "colab": {},
        "outputId": "433a69d4-432e-4564-c640-d7bf8379c6d1"
      },
      "source": [
        "# quick view on seperated data\n",
        "bogo.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>id_x</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>id_y</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward</th>\n",
              "      <th>duration</th>\n",
              "      <th>pre_offer_id</th>\n",
              "      <th>completed_offer</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>408</td>\n",
              "      <td>{'offer id': 'f19421c1d4aa40978ebb69ca19b0e20d'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>5</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5</td>\n",
              "      <td>5.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>offer completed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>414</td>\n",
              "      <td>{'offer_id': 'f19421c1d4aa40978ebb69ca19b0e20d...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>5</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5</td>\n",
              "      <td>5.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>456</td>\n",
              "      <td>{'offer id': 'f19421c1d4aa40978ebb69ca19b0e20d'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>5</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5</td>\n",
              "      <td>5.0</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>528</td>\n",
              "      <td>{'amount': 14.11}</td>\n",
              "      <td>14.11</td>\n",
              "      <td>NaN</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>5</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5</td>\n",
              "      <td>5.0</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>offer received</td>\n",
              "      <td>00116118485d4dfda04fdbaba9a87b5c</td>\n",
              "      <td>168</td>\n",
              "      <td>{'offer id': 'f19421c1d4aa40978ebb69ca19b0e20d'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>5</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5</td>\n",
              "      <td>5.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "              event                            person  time  \\\n",
              "5    offer received  0009655768c64bdeb2e877511632db8f   408   \n",
              "7   offer completed  0009655768c64bdeb2e877511632db8f   414   \n",
              "8      offer viewed  0009655768c64bdeb2e877511632db8f   456   \n",
              "10      transaction  0009655768c64bdeb2e877511632db8f   528   \n",
              "20   offer received  00116118485d4dfda04fdbaba9a87b5c   168   \n",
              "\n",
              "                                                value  amount  \\\n",
              "5    {'offer id': 'f19421c1d4aa40978ebb69ca19b0e20d'}     NaN   \n",
              "7   {'offer_id': 'f19421c1d4aa40978ebb69ca19b0e20d...     NaN   \n",
              "8    {'offer id': 'f19421c1d4aa40978ebb69ca19b0e20d'}     NaN   \n",
              "10                                  {'amount': 14.11}   14.11   \n",
              "20   {'offer id': 'f19421c1d4aa40978ebb69ca19b0e20d'}     NaN   \n",
              "\n",
              "                                id_x                          offer_id  \\\n",
              "5   f19421c1d4aa40978ebb69ca19b0e20d  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "7   f19421c1d4aa40978ebb69ca19b0e20d  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "8   f19421c1d4aa40978ebb69ca19b0e20d  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "10                               NaN  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "20  f19421c1d4aa40978ebb69ca19b0e20d  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "\n",
              "                        channels  difficulty  \\\n",
              "5   [web, email, mobile, social]           5   \n",
              "7   [web, email, mobile, social]           5   \n",
              "8   [web, email, mobile, social]           5   \n",
              "10  [web, email, mobile, social]           5   \n",
              "20  [web, email, mobile, social]           5   \n",
              "\n",
              "                                id_y offer_type  reward  duration  \\\n",
              "5   f19421c1d4aa40978ebb69ca19b0e20d       bogo       5       5.0   \n",
              "7   f19421c1d4aa40978ebb69ca19b0e20d       bogo       5       5.0   \n",
              "8   f19421c1d4aa40978ebb69ca19b0e20d       bogo       5       5.0   \n",
              "10  f19421c1d4aa40978ebb69ca19b0e20d       bogo       5       5.0   \n",
              "20  f19421c1d4aa40978ebb69ca19b0e20d       bogo       5       5.0   \n",
              "\n",
              "                        pre_offer_id  completed_offer  \n",
              "5                                NaN              NaN  \n",
              "7                                NaN              0.0  \n",
              "8   f19421c1d4aa40978ebb69ca19b0e20d              1.0  \n",
              "10  f19421c1d4aa40978ebb69ca19b0e20d              1.0  \n",
              "20                               NaN              NaN  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1rBi3alXyBI_",
        "colab_type": "code",
        "colab": {},
        "outputId": "11ccb349-f275-420d-bbff-23c84ff19bdb"
      },
      "source": [
        "discount.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>id_x</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>id_y</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward</th>\n",
              "      <th>duration</th>\n",
              "      <th>pre_offer_id</th>\n",
              "      <th>completed_offer</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>504</td>\n",
              "      <td>{'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>10</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>discount</td>\n",
              "      <td>2</td>\n",
              "      <td>10.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>offer completed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>528</td>\n",
              "      <td>{'offer_id': 'fafdcd668e3743c1bb461111dcafc2a4...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>10</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>discount</td>\n",
              "      <td>2</td>\n",
              "      <td>10.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>540</td>\n",
              "      <td>{'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>10</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>discount</td>\n",
              "      <td>2</td>\n",
              "      <td>10.0</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>552</td>\n",
              "      <td>{'amount': 13.56}</td>\n",
              "      <td>13.56</td>\n",
              "      <td>NaN</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>10</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>discount</td>\n",
              "      <td>2</td>\n",
              "      <td>10.0</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>576</td>\n",
              "      <td>{'offer id': '2906b810c7d4411798c6938adc9daaa5'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2906b810c7d4411798c6938adc9daaa5</td>\n",
              "      <td>2906b810c7d4411798c6938adc9daaa5</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>10</td>\n",
              "      <td>2906b810c7d4411798c6938adc9daaa5</td>\n",
              "      <td>discount</td>\n",
              "      <td>2</td>\n",
              "      <td>7.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "              event                            person  time  \\\n",
              "9    offer received  0009655768c64bdeb2e877511632db8f   504   \n",
              "11  offer completed  0009655768c64bdeb2e877511632db8f   528   \n",
              "12     offer viewed  0009655768c64bdeb2e877511632db8f   540   \n",
              "13      transaction  0009655768c64bdeb2e877511632db8f   552   \n",
              "14   offer received  0009655768c64bdeb2e877511632db8f   576   \n",
              "\n",
              "                                                value  amount  \\\n",
              "9    {'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}     NaN   \n",
              "11  {'offer_id': 'fafdcd668e3743c1bb461111dcafc2a4...     NaN   \n",
              "12   {'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}     NaN   \n",
              "13                                  {'amount': 13.56}   13.56   \n",
              "14   {'offer id': '2906b810c7d4411798c6938adc9daaa5'}     NaN   \n",
              "\n",
              "                                id_x                          offer_id  \\\n",
              "9   fafdcd668e3743c1bb461111dcafc2a4  fafdcd668e3743c1bb461111dcafc2a4   \n",
              "11  fafdcd668e3743c1bb461111dcafc2a4  fafdcd668e3743c1bb461111dcafc2a4   \n",
              "12  fafdcd668e3743c1bb461111dcafc2a4  fafdcd668e3743c1bb461111dcafc2a4   \n",
              "13                               NaN  fafdcd668e3743c1bb461111dcafc2a4   \n",
              "14  2906b810c7d4411798c6938adc9daaa5  2906b810c7d4411798c6938adc9daaa5   \n",
              "\n",
              "                        channels  difficulty  \\\n",
              "9   [web, email, mobile, social]          10   \n",
              "11  [web, email, mobile, social]          10   \n",
              "12  [web, email, mobile, social]          10   \n",
              "13  [web, email, mobile, social]          10   \n",
              "14          [web, email, mobile]          10   \n",
              "\n",
              "                                id_y offer_type  reward  duration  \\\n",
              "9   fafdcd668e3743c1bb461111dcafc2a4   discount       2      10.0   \n",
              "11  fafdcd668e3743c1bb461111dcafc2a4   discount       2      10.0   \n",
              "12  fafdcd668e3743c1bb461111dcafc2a4   discount       2      10.0   \n",
              "13  fafdcd668e3743c1bb461111dcafc2a4   discount       2      10.0   \n",
              "14  2906b810c7d4411798c6938adc9daaa5   discount       2       7.0   \n",
              "\n",
              "                        pre_offer_id  completed_offer  \n",
              "9                                NaN              NaN  \n",
              "11                               NaN              0.0  \n",
              "12  fafdcd668e3743c1bb461111dcafc2a4              1.0  \n",
              "13  fafdcd668e3743c1bb461111dcafc2a4              1.0  \n",
              "14                               NaN              NaN  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qRRKGkVPyBJF",
        "colab_type": "code",
        "colab": {},
        "outputId": "d83480e1-7d94-4409-8efb-014c2de85e2e"
      },
      "source": [
        "informational.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>id_x</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>id_y</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward</th>\n",
              "      <th>duration</th>\n",
              "      <th>pre_offer_id</th>\n",
              "      <th>completed_offer</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>168</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>192</td>\n",
              "      <td>{'offer id': '5a8bc65990b245e5a138643cd4eb9837'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>228</td>\n",
              "      <td>{'amount': 22.16}</td>\n",
              "      <td>22.16</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>336</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>372</td>\n",
              "      <td>{'offer id': '3f207df678b143eea3cee63160fa8bed'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            event                            person  time  \\\n",
              "0  offer received  0009655768c64bdeb2e877511632db8f   168   \n",
              "1    offer viewed  0009655768c64bdeb2e877511632db8f   192   \n",
              "2     transaction  0009655768c64bdeb2e877511632db8f   228   \n",
              "3  offer received  0009655768c64bdeb2e877511632db8f   336   \n",
              "4    offer viewed  0009655768c64bdeb2e877511632db8f   372   \n",
              "\n",
              "                                              value  amount  \\\n",
              "0  {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "1  {'offer id': '5a8bc65990b245e5a138643cd4eb9837'}     NaN   \n",
              "2                                 {'amount': 22.16}   22.16   \n",
              "3  {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "4  {'offer id': '3f207df678b143eea3cee63160fa8bed'}     NaN   \n",
              "\n",
              "                               id_x                          offer_id  \\\n",
              "0  5a8bc65990b245e5a138643cd4eb9837  5a8bc65990b245e5a138643cd4eb9837   \n",
              "1  5a8bc65990b245e5a138643cd4eb9837  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2                               NaN  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3  3f207df678b143eea3cee63160fa8bed  3f207df678b143eea3cee63160fa8bed   \n",
              "4  3f207df678b143eea3cee63160fa8bed  3f207df678b143eea3cee63160fa8bed   \n",
              "\n",
              "                  channels  difficulty                              id_y  \\\n",
              "0  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "1  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3     [web, email, mobile]           0  3f207df678b143eea3cee63160fa8bed   \n",
              "4     [web, email, mobile]           0  3f207df678b143eea3cee63160fa8bed   \n",
              "\n",
              "      offer_type  reward  duration                      pre_offer_id  \\\n",
              "0  informational       0       3.0                               NaN   \n",
              "1  informational       0       3.0                               NaN   \n",
              "2  informational       0       3.0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3  informational       0       4.0                               NaN   \n",
              "4  informational       0       4.0                               NaN   \n",
              "\n",
              "   completed_offer  \n",
              "0              NaN  \n",
              "1              0.0  \n",
              "2              1.0  \n",
              "3              NaN  \n",
              "4              0.0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GZi_HoP-yBJJ",
        "colab_type": "text"
      },
      "source": [
        "Within each offer type, use responded_offer flagged in previous steps we can filter out the offers which were successfully viewed, and completed by users. For BOGO and discount offer, the responsed offer should be the one that with 'offer complete' events, and for the informational offer, just 'transaction' can be seen as a successful offer."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LlURXOYnyBJJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# extract responded offer under bogo and informational type\n",
        "bogo_completed = bogo[['person','offer_id']][(bogo['completed_offer']==1) & (bogo['event']=='offer completed')].groupby(['person','offer_id']).count().reset_index()\n",
        "discount_completed = discount[['person','offer_id']][(discount['completed_offer']==1) & (discount['event']=='offer completed')].groupby(['person','offer_id']).count().reset_index()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CRb69Ow-yBJL",
        "colab_type": "text"
      },
      "source": [
        "Next, will seperate out customers who only viewed the offers without transaction and completion in the end, and the customers who only received the offer without viewing it."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jfJt00RYyBJM",
        "colab_type": "code",
        "colab": {},
        "outputId": "fc5c46d5-63b4-47c3-c6cb-9ec1ca96bae0"
      },
      "source": [
        "# filter out offer with transactions or completed, and offer which have viewed events\n",
        "bogo_ids_transaction_completed = bogo[['person','offer_id']][(bogo['event']=='transaction') | \\\n",
        "                            (bogo['event']=='offer completed') ].groupby(['person','offer_id']).count().reset_index()\n",
        "bogo_ids_received = bogo[['person','offer_id']][bogo['event']=='offer received'].groupby(['person','offer_id']).count().reset_index()\n",
        "\n",
        "# get the offer records which was only viewed without transaction and completion\n",
        "bogo_merged = bogo_ids_transaction_completed.merge(bogo_ids_received,how='right',on=['person','offer_id'],indicator=True)\n",
        "bogo_merged.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>person</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>_merge</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>00116118485d4dfda04fdbaba9a87b5c</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0011e0d4e6b944f998e987f904e8c1e5</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0020c2b971eb4e9188eac86d93036a77</td>\n",
              "      <td>4d5c57ea9a6940dd891ad53e9dbe8da0</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             person                          offer_id _merge\n",
              "0  0009655768c64bdeb2e877511632db8f  f19421c1d4aa40978ebb69ca19b0e20d   both\n",
              "1  00116118485d4dfda04fdbaba9a87b5c  f19421c1d4aa40978ebb69ca19b0e20d   both\n",
              "2  0011e0d4e6b944f998e987f904e8c1e5  9b98b8c7a33c4b65b9aebfe6a799e6d9   both\n",
              "3  0020c2b971eb4e9188eac86d93036a77  4d5c57ea9a6940dd891ad53e9dbe8da0   both\n",
              "4  0020ccbbb6d84e358d3414a3ff76cffd  9b98b8c7a33c4b65b9aebfe6a799e6d9   both"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "baauNHDnyBJP",
        "colab_type": "code",
        "colab": {},
        "outputId": "aaf8b9e1-f4c1-4c1a-fb84-c1466de47977"
      },
      "source": [
        "# do the same processing on discountrmational offer\n",
        "discount_ids_transaction_completed = discount[['person','offer_id']][(discount['event']=='transaction') | \\\n",
        "                            (discount['event']=='offer completed') ].groupby(['person','offer_id']).count().reset_index()\n",
        "discount_ids_received = discount[['person','offer_id']][discount['event']=='offer received'].groupby(['person','offer_id']).count().reset_index()\n",
        "\n",
        "# get the offer records which was only viewed without transaction and completion\n",
        "discount_merged = discount_ids_transaction_completed.merge(discount_ids_received,how='right',on=['person','offer_id'],indicator=True)\n",
        "discount_merged.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>person</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>_merge</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>2906b810c7d4411798c6938adc9daaa5</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0011e0d4e6b944f998e987f904e8c1e5</td>\n",
              "      <td>0b1e1539f2cc45b7b9fa7c272da2e1d7</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0011e0d4e6b944f998e987f904e8c1e5</td>\n",
              "      <td>2298d6c36e964ae4a3e7e9706d1fb8c2</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0020c2b971eb4e9188eac86d93036a77</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             person                          offer_id _merge\n",
              "0  0009655768c64bdeb2e877511632db8f  2906b810c7d4411798c6938adc9daaa5   both\n",
              "1  0009655768c64bdeb2e877511632db8f  fafdcd668e3743c1bb461111dcafc2a4   both\n",
              "2  0011e0d4e6b944f998e987f904e8c1e5  0b1e1539f2cc45b7b9fa7c272da2e1d7   both\n",
              "3  0011e0d4e6b944f998e987f904e8c1e5  2298d6c36e964ae4a3e7e9706d1fb8c2   both\n",
              "4  0020c2b971eb4e9188eac86d93036a77  fafdcd668e3743c1bb461111dcafc2a4   both"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AY1eXZgsyBJT",
        "colab_type": "text"
      },
      "source": [
        "Then, based on merged dataset above, we can seperate out customers who only viewed the offer after they received the offer and customers who didn't even open the offer after they receive the offer."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IHG1nhbnyBJU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# subtract the customers who received the offer without finishing the transaction\n",
        "bogo_no_transaction = bogo_merged[bogo_merged['_merge']=='right_only']\n",
        "bogo_no_transaction = bogo_no_transaction.merge(transcript_processed, how='left',on=['person','offer_id'])\n",
        "\n",
        "# subset users who only veiwed the offer and people who viewed the offer but didn't make transaction\n",
        "bogo_viewed = bogo_no_transaction[['person','offer_id']][bogo_no_transaction['event']=='offer viewed']\\\n",
        "                .groupby(['person','offer_id']).count().reset_index()\n",
        "bogo_no_transaction.drop(['_merge'], axis=1, inplace=True)\n",
        "\n",
        "# get the customers who didn't even view the offer after they received the BOGO offer \n",
        "bogo_no_view = bogo_no_transaction.merge(bogo_viewed[['person','offer_id']], how='left', indicator=True)\n",
        "bogo_no_view = bogo_no_view[bogo_no_view['_merge']=='left_only'].copy()\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "O4U6yRPEyBJW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# do the same action on discountrmational offer\n",
        "\n",
        "# subtract the customers who received the offer without finishing the transaction\n",
        "discount_no_transaction = discount_merged[discount_merged['_merge']=='right_only']\n",
        "discount_no_transaction = discount_no_transaction.merge(transcript_processed, how='left',on=['person','offer_id'])\n",
        "\n",
        "# subset users who only veiwed the offer and people who viewed the offer but didn't make transaction\n",
        "discount_viewed = discount_no_transaction[['person','offer_id']][discount_no_transaction['event']=='offer viewed']\\\n",
        "                .groupby(['person','offer_id']).count().reset_index()\n",
        "discount_no_transaction.drop(['_merge'], axis=1, inplace=True)\n",
        "\n",
        "# get the customers who didn't even view the offer after they received the discount offer \n",
        "discount_no_view = discount_no_transaction.merge(discount_viewed[['person','offer_id']], how='left', indicator=True)\n",
        "discount_no_view = discount_no_view[discount_no_view['_merge']=='left_only'].copy()\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JUbE64CFyBJZ",
        "colab_type": "text"
      },
      "source": [
        "After above processing, filter out the transaction regardless of receiving or viewing the offer."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YSEbTB4PyBJa",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# subset the offer which has no correction with offer\n",
        "bogo_unrelated = bogo[['person','offer_id']][(bogo['event']=='offer completed') & (bogo['completed_offer']!=1)]\\\n",
        "                    .groupby(['person','offer_id']).count().reset_index()\n",
        "discount_unrelated = discount[['person','offer_id']][(discount['event']=='offer completed')\\\n",
        "                    & (discount['completed_offer']!=1)].groupby(['person','offer_id']).count().reset_index()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AMpqB3SWyBJc",
        "colab_type": "text"
      },
      "source": [
        "After separating the different cases of customers, the following steps will firstly focus on customers who finish the transaction after receiving the offer and customers who only view the offer without any transaction."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BZlozjL_yBJc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# combine the two kind of customers cases which are focused on\n",
        "bogo_completed['offer_responded']=1\n",
        "bogo_viewed['offer_responded']=0\n",
        "bogo_offer = bogo_completed.append(bogo_viewed, sort=False)\n",
        "\n",
        "discount_completed['offer_responded']=1\n",
        "discount_viewed['offer_responded']=0\n",
        "discount_offer = discount_completed.append(discount_viewed, sort=False)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1Zfs2OnCyBJg",
        "colab_type": "text"
      },
      "source": [
        "As for the informational offer, the offer could only be counted as responded under the effect of the offer when the transaction is finished within the duration of the offer."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NvktCn5HyBJg",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#convert time into days\n",
        "informational['day_offer']=informational['time']/24\n",
        "#drop unnecessary columns\n",
        "informational.drop(['time','value','pre_offer_id'], axis=1, inplace=True)\n",
        "informational=informational.sort_values(['person','day_offer','event','offer_id'])\n",
        "\n",
        "#get difference in time for informational offers\n",
        "informational['diff']=informational[(informational['offer_type']=='informational') \\\n",
        "                                         & ((informational['event']=='offer received')\\\n",
        "                                            | (informational['event']=='transaction'))].groupby(['person','offer_id'])['day_offer'].diff()\n",
        "informational['valid_completion']=np.nan\n",
        "#flag valid events if within duration\n",
        "informational.loc[informational['diff']<=informational['duration'],'valid_completion']=1\n",
        "#fill any missing values with 0 flag\n",
        "informational['valid_completion']=informational['valid_completion'].fillna(value=0)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WGhlvr6CyBJm",
        "colab_type": "code",
        "colab": {},
        "outputId": "e9dfd32b-2ebd-4aee-821f-40c02803ea7b"
      },
      "source": [
        "# quick check on dataset\n",
        "informational.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>amount</th>\n",
              "      <th>id_x</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>id_y</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward</th>\n",
              "      <th>duration</th>\n",
              "      <th>completed_offer</th>\n",
              "      <th>day_offer</th>\n",
              "      <th>diff</th>\n",
              "      <th>valid_completion</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>7.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>8.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>transaction</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>22.16</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>[email, mobile, social]</td>\n",
              "      <td>0</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>9.5</td>\n",
              "      <td>2.5</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>offer received</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>14.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer viewed</td>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>15.5</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            event                            person  amount  \\\n",
              "0  offer received  0009655768c64bdeb2e877511632db8f     NaN   \n",
              "1    offer viewed  0009655768c64bdeb2e877511632db8f     NaN   \n",
              "2     transaction  0009655768c64bdeb2e877511632db8f   22.16   \n",
              "3  offer received  0009655768c64bdeb2e877511632db8f     NaN   \n",
              "4    offer viewed  0009655768c64bdeb2e877511632db8f     NaN   \n",
              "\n",
              "                               id_x                          offer_id  \\\n",
              "0  5a8bc65990b245e5a138643cd4eb9837  5a8bc65990b245e5a138643cd4eb9837   \n",
              "1  5a8bc65990b245e5a138643cd4eb9837  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2                               NaN  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3  3f207df678b143eea3cee63160fa8bed  3f207df678b143eea3cee63160fa8bed   \n",
              "4  3f207df678b143eea3cee63160fa8bed  3f207df678b143eea3cee63160fa8bed   \n",
              "\n",
              "                  channels  difficulty                              id_y  \\\n",
              "0  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "1  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2  [email, mobile, social]           0  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3     [web, email, mobile]           0  3f207df678b143eea3cee63160fa8bed   \n",
              "4     [web, email, mobile]           0  3f207df678b143eea3cee63160fa8bed   \n",
              "\n",
              "      offer_type  reward  duration  completed_offer  day_offer  diff  \\\n",
              "0  informational       0       3.0              NaN        7.0   NaN   \n",
              "1  informational       0       3.0              0.0        8.0   NaN   \n",
              "2  informational       0       3.0              1.0        9.5   2.5   \n",
              "3  informational       0       4.0              NaN       14.0   NaN   \n",
              "4  informational       0       4.0              0.0       15.5   NaN   \n",
              "\n",
              "   valid_completion  \n",
              "0               0.0  \n",
              "1               0.0  \n",
              "2               1.0  \n",
              "3               0.0  \n",
              "4               0.0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_u6VOIaxyBJn",
        "colab_type": "text"
      },
      "source": [
        "As shown above, only when customer make the transaction during the offer effective duration, it can be counted as successfully responsed offer."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zJPsMBssyBJn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#flag successfully responded offer where completed_offer=1 and valid_completion=1\n",
        "informational['offer_responded']=np.where(((informational['completed_offer']==1)\\\n",
        "                                           & (informational['valid_completion']==1)),1,0)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-jCmTQ1YyBJp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# extract the informational offer successfully compelted after\n",
        "info_completed = informational[['person','offer_id']][informational['offer_responded']==1]\\\n",
        "                        .groupby(['person','offer_id']).sum().reset_index()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iZUKKd3JyBJr",
        "colab_type": "code",
        "colab": {},
        "outputId": "7d4e5e59-72e2-4886-f91b-92d21e2efceb"
      },
      "source": [
        "# filter out offer with transactions or completed, and offer which have viewed events\n",
        "informational_ids_transaction_completed = informational[['person','offer_id']][(informational['event']=='transaction') | \\\n",
        "                            (informational['event']=='offer completed') ].groupby(['person','offer_id']).count().reset_index()\n",
        "informational_ids_received = informational[['person','offer_id']][informational['event']=='offer received'].groupby(['person','offer_id']).count().reset_index()\n",
        "\n",
        "# get the offer records which was only viewed without transaction and completion\n",
        "informational_merged = informational_ids_transaction_completed.merge(informational_ids_received,how='right',on=['person','offer_id'],indicator=True)\n",
        "informational_merged.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>person</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>_merge</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0011e0d4e6b944f998e987f904e8c1e5</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0020c2b971eb4e9188eac86d93036a77</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>both</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             person                          offer_id _merge\n",
              "0  0009655768c64bdeb2e877511632db8f  3f207df678b143eea3cee63160fa8bed   both\n",
              "1  0009655768c64bdeb2e877511632db8f  5a8bc65990b245e5a138643cd4eb9837   both\n",
              "2  0011e0d4e6b944f998e987f904e8c1e5  3f207df678b143eea3cee63160fa8bed   both\n",
              "3  0020c2b971eb4e9188eac86d93036a77  5a8bc65990b245e5a138643cd4eb9837   both\n",
              "4  0020ccbbb6d84e358d3414a3ff76cffd  5a8bc65990b245e5a138643cd4eb9837   both"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 47
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4c6d4PRIyBJt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# subtract the customers who received the offer without finishing the transaction\n",
        "informational_no_transaction = informational_merged[informational_merged['_merge']=='right_only']\n",
        "informational_no_transaction = informational_no_transaction.merge(transcript_processed, how='left',on=['person','offer_id'])\n",
        "\n",
        "# subset users who only veiwed the offer and people who viewed the offer but didn't make transaction\n",
        "informational_viewed = informational_no_transaction[['person','offer_id']][informational_no_transaction['event']=='offer viewed']\\\n",
        "                .groupby(['person','offer_id']).count().reset_index()\n",
        "informational_no_transaction.drop(['_merge'], axis=1, inplace=True)\n",
        "\n",
        "# get the customers who didn't even view the offer after they received the informational offer \n",
        "informational_no_view = informational_no_transaction.merge(informational_viewed[['person','offer_id']], how='left', indicator=True)\n",
        "informational_no_view = informational_no_view[informational_no_view['_merge']=='left_only'].copy()\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kESphycyyBJw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# append the informational offer together\n",
        "info_completed['offer_responded']=1\n",
        "informational_viewed['offer_responded']=0\n",
        "informational_offer = info_completed.append(informational_viewed, sort=False)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NUbf7yb0yBJ0",
        "colab_type": "text"
      },
      "source": [
        "## Feature engineering\n",
        "\n",
        "After basic processing, the next step will look if there are any columns that can be used to create new features."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VWZA5wtJyBJ1",
        "colab_type": "text"
      },
      "source": [
        "- generate a new column for length of customer's membership"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KWEyZY-5yBJ2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "profile.rename(columns={'id':'person'},inplace=True)\n",
        "\n",
        "#create function to generate the length of time since the user become the member\n",
        "def member_since(df):\n",
        "    '''\n",
        "    inputs:\n",
        "    - df: dataframe to generate the ength of time since the user become the member\n",
        "    outputs:\n",
        "    - Returns dataframe with became_member_on column transformed to be tenure in days\n",
        "    \n",
        "    '''\n",
        "    #merge to get user demographic profile\n",
        "    df=df.merge(profile,how='left',on='person')\n",
        "    \n",
        "    df['year']=pd.Series([int(str(x)[:4]) for x in df['became_member_on']])\n",
        "    df['month']=pd.Series([int(str(x)[-3]) for x in df['became_member_on']])\n",
        "    df['day']=pd.Series([int(str(x)[-2:]) for x in df['became_member_on']])\n",
        "    df = df.drop('became_member_on', axis=1, inplace=False)\n",
        "    df.loc[df['year'] == 2018, 'membership_days'] = (30*df['month'])+df['day']\n",
        "    df.loc[df['year'] != 2018, 'membership_days'] = ((2018-df['year'])*365)+(30*df['month'])+df['day']\n",
        "    df = df.drop(['year','month','day'], axis=1, inplace=False)\n",
        "    \n",
        "    return df\n",
        "\n",
        "bogo_offer=member_since(bogo_offer)\n",
        "discount_offer=member_since(discount_offer)\n",
        "informational_offer=member_since(informational_offer)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ESaHj6YZyBJ6",
        "colab_type": "code",
        "colab": {},
        "outputId": "7fff7781-5992-43f1-f621-7b49544db428"
      },
      "source": [
        "bogo_offer.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>person</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>offer_responded</th>\n",
              "      <th>age</th>\n",
              "      <th>gender</th>\n",
              "      <th>income</th>\n",
              "      <th>membership_days</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0011e0d4e6b944f998e987f904e8c1e5</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>1</td>\n",
              "      <td>40</td>\n",
              "      <td>O</td>\n",
              "      <td>57000.0</td>\n",
              "      <td>39.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0020c2b971eb4e9188eac86d93036a77</td>\n",
              "      <td>4d5c57ea9a6940dd891ad53e9dbe8da0</td>\n",
              "      <td>1</td>\n",
              "      <td>59</td>\n",
              "      <td>F</td>\n",
              "      <td>90000.0</td>\n",
              "      <td>824.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>F</td>\n",
              "      <td>60000.0</td>\n",
              "      <td>771.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>F</td>\n",
              "      <td>60000.0</td>\n",
              "      <td>771.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>004b041fbfe44859945daa2c7f79ee64</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>1</td>\n",
              "      <td>55</td>\n",
              "      <td>F</td>\n",
              "      <td>74000.0</td>\n",
              "      <td>158.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             person                          offer_id  \\\n",
              "0  0011e0d4e6b944f998e987f904e8c1e5  9b98b8c7a33c4b65b9aebfe6a799e6d9   \n",
              "1  0020c2b971eb4e9188eac86d93036a77  4d5c57ea9a6940dd891ad53e9dbe8da0   \n",
              "2  0020ccbbb6d84e358d3414a3ff76cffd  9b98b8c7a33c4b65b9aebfe6a799e6d9   \n",
              "3  0020ccbbb6d84e358d3414a3ff76cffd  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "4  004b041fbfe44859945daa2c7f79ee64  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "\n",
              "   offer_responded  age gender   income  membership_days  \n",
              "0                1   40      O  57000.0             39.0  \n",
              "1                1   59      F  90000.0            824.0  \n",
              "2                1   24      F  60000.0            771.0  \n",
              "3                1   24      F  60000.0            771.0  \n",
              "4                1   55      F  74000.0            158.0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D736MGgryBJ9",
        "colab_type": "text"
      },
      "source": [
        "- generate the count of offer received for each user"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1XU793EIyBJ9",
        "colab_type": "code",
        "colab": {},
        "outputId": "b7f6891f-33ce-4d1b-9d2f-687349e5e267"
      },
      "source": [
        "transcript.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>event</th>\n",
              "      <th>person</th>\n",
              "      <th>time</th>\n",
              "      <th>value</th>\n",
              "      <th>amount</th>\n",
              "      <th>reward_x</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>channels</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>duration</th>\n",
              "      <th>id</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward_y</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>offer received</td>\n",
              "      <td>78afa995795e4d85b5d9ceeca43f5fef</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '9b98b8c7a33c4b65b9aebfe6a799e6d9'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>5.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>offer received</td>\n",
              "      <td>a03223e636434f42ac4c3df47e8bac43</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '0b1e1539f2cc45b7b9fa7c272da2e1d7'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0b1e1539f2cc45b7b9fa7c272da2e1d7</td>\n",
              "      <td>[web, email]</td>\n",
              "      <td>20.0</td>\n",
              "      <td>10.0</td>\n",
              "      <td>0b1e1539f2cc45b7b9fa7c272da2e1d7</td>\n",
              "      <td>discount</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>offer received</td>\n",
              "      <td>e2127556f4f64592b11af22de27a7932</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '2906b810c7d4411798c6938adc9daaa5'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2906b810c7d4411798c6938adc9daaa5</td>\n",
              "      <td>[web, email, mobile]</td>\n",
              "      <td>10.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>2906b810c7d4411798c6938adc9daaa5</td>\n",
              "      <td>discount</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>offer received</td>\n",
              "      <td>8ec6ce2a7e7949b1bf142def7d0e0586</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>10.0</td>\n",
              "      <td>10.0</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>discount</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>offer received</td>\n",
              "      <td>68617ca6246f4fbc85e91a2a49552598</td>\n",
              "      <td>0</td>\n",
              "      <td>{'offer id': '4d5c57ea9a6940dd891ad53e9dbe8da0'}</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>4d5c57ea9a6940dd891ad53e9dbe8da0</td>\n",
              "      <td>[web, email, mobile, social]</td>\n",
              "      <td>10.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>4d5c57ea9a6940dd891ad53e9dbe8da0</td>\n",
              "      <td>bogo</td>\n",
              "      <td>10.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            event                            person  time  \\\n",
              "0  offer received  78afa995795e4d85b5d9ceeca43f5fef     0   \n",
              "1  offer received  a03223e636434f42ac4c3df47e8bac43     0   \n",
              "2  offer received  e2127556f4f64592b11af22de27a7932     0   \n",
              "3  offer received  8ec6ce2a7e7949b1bf142def7d0e0586     0   \n",
              "4  offer received  68617ca6246f4fbc85e91a2a49552598     0   \n",
              "\n",
              "                                              value  amount  reward_x  \\\n",
              "0  {'offer id': '9b98b8c7a33c4b65b9aebfe6a799e6d9'}     NaN       NaN   \n",
              "1  {'offer id': '0b1e1539f2cc45b7b9fa7c272da2e1d7'}     NaN       NaN   \n",
              "2  {'offer id': '2906b810c7d4411798c6938adc9daaa5'}     NaN       NaN   \n",
              "3  {'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}     NaN       NaN   \n",
              "4  {'offer id': '4d5c57ea9a6940dd891ad53e9dbe8da0'}     NaN       NaN   \n",
              "\n",
              "                           offer_id                      channels  difficulty  \\\n",
              "0  9b98b8c7a33c4b65b9aebfe6a799e6d9          [web, email, mobile]         5.0   \n",
              "1  0b1e1539f2cc45b7b9fa7c272da2e1d7                  [web, email]        20.0   \n",
              "2  2906b810c7d4411798c6938adc9daaa5          [web, email, mobile]        10.0   \n",
              "3  fafdcd668e3743c1bb461111dcafc2a4  [web, email, mobile, social]        10.0   \n",
              "4  4d5c57ea9a6940dd891ad53e9dbe8da0  [web, email, mobile, social]        10.0   \n",
              "\n",
              "   duration                                id offer_type  reward_y  \n",
              "0       7.0  9b98b8c7a33c4b65b9aebfe6a799e6d9       bogo       5.0  \n",
              "1      10.0  0b1e1539f2cc45b7b9fa7c272da2e1d7   discount       5.0  \n",
              "2       7.0  2906b810c7d4411798c6938adc9daaa5   discount       2.0  \n",
              "3      10.0  fafdcd668e3743c1bb461111dcafc2a4   discount       2.0  \n",
              "4       5.0  4d5c57ea9a6940dd891ad53e9dbe8da0       bogo      10.0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": true,
        "id": "8jFvDHQuyBJ_",
        "colab_type": "code",
        "colab": {},
        "outputId": "e5276b74-aeaa-4cb2-b3aa-9a3bd4e40b30"
      },
      "source": [
        "# generate the count of offers received per person\n",
        "offer_cnt=transcript_processed[transcript_processed['event']=='offer received'].groupby(['person','offer_id','time']).count()['event'].reset_index()\n",
        "offer_cnt.rename(columns={'event':'offer_received_cnt'},inplace=True)\n",
        "\n",
        "offer_cnt.drop(['time'], axis=1, inplace=True)\n",
        "\n",
        "# ensure only unique person-offer_id pairs\n",
        "offer_cnt=offer_cnt.groupby(['person','offer_id']).sum().reset_index()\n",
        "offer_cnt.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>person</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>offer_received_cnt</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>2906b810c7d4411798c6938adc9daaa5</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             person                          offer_id  \\\n",
              "0  0009655768c64bdeb2e877511632db8f  2906b810c7d4411798c6938adc9daaa5   \n",
              "1  0009655768c64bdeb2e877511632db8f  3f207df678b143eea3cee63160fa8bed   \n",
              "2  0009655768c64bdeb2e877511632db8f  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3  0009655768c64bdeb2e877511632db8f  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "4  0009655768c64bdeb2e877511632db8f  fafdcd668e3743c1bb461111dcafc2a4   \n",
              "\n",
              "   offer_received_cnt  \n",
              "0                   1  \n",
              "1                   1  \n",
              "2                   1  \n",
              "3                   1  \n",
              "4                   1  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 53
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MyB3iBXiyBKB",
        "colab_type": "text"
      },
      "source": [
        "- subtract the transactions which's not related to the offer"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vd4vwc6syBKC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# subtract the transactions which's not related to the offer\n",
        "transactions_not_related=transcript_processed[(transcript_processed['event']=='transaction') & (transcript_processed['completed_offer']==0)].groupby(['person','offer_id'])['amount'].sum().reset_index()\n",
        "transactions_not_related.rename(columns={'amount':'amount_invalid'},inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "J1x2o-akyBKI",
        "colab_type": "text"
      },
      "source": [
        "- calculate the time lap between offers"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nHthGRYCyBKJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# convert time into days\n",
        "transcript_processed['day_offer']=transcript_processed['time']/24\n",
        "# drop unnecessary columns\n",
        "transcript_processed.drop(['time'], axis=1, inplace=True)\n",
        "\n",
        "# calculate the time between offers\n",
        "transcript_processed['time_gap']=transcript_processed[transcript_processed['event']=='offer received'].groupby(['person','offer_id'])['day_offer'].diff()\n",
        "\n",
        "# fill missing values with 0\n",
        "transcript_processed['time_gap']=transcript_processed['time_gap'].fillna(value=0)\n",
        "\n",
        "df_time_gap=transcript_processed.groupby(['person','offer_id'])['time_gap'].sum().reset_index()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "blILZhnAyBKM",
        "colab_type": "text"
      },
      "source": [
        "- Merge the temporary data created above together, then drop the missing values in gender column, and split the channel column to the categorical variable"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PWAf9t5myBKM",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# merge to get offers received count and invalid amount transacted \n",
        "bogo_offer=bogo_offer.merge(offer_cnt[['person','offer_id','offer_received_cnt']],how='left',on=['person','offer_id'])\n",
        "bogo_offer=bogo_offer.merge(transactions_not_related[['person','offer_id','amount_invalid']],how='left',on=['person','offer_id'])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PJlceET1yBKO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# fill missing values for amount_invalid with 0\n",
        "bogo_offer['amount_invalid']=bogo_offer['amount_invalid'].fillna(value=0)\n",
        "bogo_offer.dropna(inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BnIHmu4FyBKQ",
        "colab_type": "code",
        "colab": {},
        "outputId": "82b981dc-e6fb-4d86-8faf-46199bfcb023"
      },
      "source": [
        "bogo_offer.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>person</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>offer_responded</th>\n",
              "      <th>age</th>\n",
              "      <th>gender</th>\n",
              "      <th>income</th>\n",
              "      <th>membership_days</th>\n",
              "      <th>offer_received_cnt</th>\n",
              "      <th>amount_invalid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0011e0d4e6b944f998e987f904e8c1e5</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>1</td>\n",
              "      <td>40</td>\n",
              "      <td>O</td>\n",
              "      <td>57000.0</td>\n",
              "      <td>39.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0020c2b971eb4e9188eac86d93036a77</td>\n",
              "      <td>4d5c57ea9a6940dd891ad53e9dbe8da0</td>\n",
              "      <td>1</td>\n",
              "      <td>59</td>\n",
              "      <td>F</td>\n",
              "      <td>90000.0</td>\n",
              "      <td>824.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>F</td>\n",
              "      <td>60000.0</td>\n",
              "      <td>771.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>F</td>\n",
              "      <td>60000.0</td>\n",
              "      <td>771.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>004b041fbfe44859945daa2c7f79ee64</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>1</td>\n",
              "      <td>55</td>\n",
              "      <td>F</td>\n",
              "      <td>74000.0</td>\n",
              "      <td>158.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             person                          offer_id  \\\n",
              "0  0011e0d4e6b944f998e987f904e8c1e5  9b98b8c7a33c4b65b9aebfe6a799e6d9   \n",
              "1  0020c2b971eb4e9188eac86d93036a77  4d5c57ea9a6940dd891ad53e9dbe8da0   \n",
              "2  0020ccbbb6d84e358d3414a3ff76cffd  9b98b8c7a33c4b65b9aebfe6a799e6d9   \n",
              "3  0020ccbbb6d84e358d3414a3ff76cffd  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "4  004b041fbfe44859945daa2c7f79ee64  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "\n",
              "   offer_responded  age gender   income  membership_days  offer_received_cnt  \\\n",
              "0                1   40      O  57000.0             39.0                   1   \n",
              "1                1   59      F  90000.0            824.0                   1   \n",
              "2                1   24      F  60000.0            771.0                   1   \n",
              "3                1   24      F  60000.0            771.0                   1   \n",
              "4                1   55      F  74000.0            158.0                   1   \n",
              "\n",
              "   amount_invalid  \n",
              "0             0.0  \n",
              "1             0.0  \n",
              "2             0.0  \n",
              "3             0.0  \n",
              "4             0.0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 58
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NKkoYLmVyBKS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Split the channel column into categorical variables\n",
        "def rename(col_name,df):\n",
        "    df[col_name]=np.where(df[col_name]==col_name,1,0)\n",
        "    return df\n",
        "\n",
        "# dummy variable encoding function\n",
        "def dummy(df,col):\n",
        "    df = pd.concat([df[:],pd.get_dummies(df[col],prefix=col)],axis=1)\n",
        "    df = df.drop([col], axis=1, inplace=False)\n",
        "    return df\n",
        "\n",
        "portfolio.rename(columns={'id':'offer_id'}, inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X9d4LsaGyBKV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# merge with portfolio to get offer details\n",
        "bogo_offer=bogo_offer.merge(portfolio,how='left',on='offer_id')\n",
        "\n",
        "# convert channels into categorical variables\n",
        "channels = bogo_offer['channels'].apply(pd.Series)\n",
        "channels = channels.rename(columns={0:'web',1:'email',2:'mobile',3:'social'})\n",
        "bogo_offer = pd.concat([bogo_offer[:], channels[:]], axis=1)\n",
        "rename('web',bogo_offer)\n",
        "rename('email',bogo_offer)\n",
        "rename('mobile',bogo_offer)\n",
        "rename('social',bogo_offer)\n",
        "bogo_offer = bogo_offer.drop(['channels'], axis=1, inplace=False)\n",
        "\n",
        "# convert gender into categorical variables\n",
        "bogo_offer=dummy(bogo_offer,'gender')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ev7EZ_k_yBKW",
        "colab_type": "code",
        "colab": {},
        "outputId": "db74ddfd-d51e-4fd1-b58f-19655bd19309"
      },
      "source": [
        "# quick check on processed data\n",
        "bogo_offer.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>person</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>offer_responded</th>\n",
              "      <th>age</th>\n",
              "      <th>income</th>\n",
              "      <th>membership_days</th>\n",
              "      <th>offer_received_cnt</th>\n",
              "      <th>amount_invalid</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>duration</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward</th>\n",
              "      <th>web</th>\n",
              "      <th>email</th>\n",
              "      <th>mobile</th>\n",
              "      <th>social</th>\n",
              "      <th>gender_F</th>\n",
              "      <th>gender_M</th>\n",
              "      <th>gender_O</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0011e0d4e6b944f998e987f904e8c1e5</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>1</td>\n",
              "      <td>40</td>\n",
              "      <td>57000.0</td>\n",
              "      <td>39.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5</td>\n",
              "      <td>7</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0020c2b971eb4e9188eac86d93036a77</td>\n",
              "      <td>4d5c57ea9a6940dd891ad53e9dbe8da0</td>\n",
              "      <td>1</td>\n",
              "      <td>59</td>\n",
              "      <td>90000.0</td>\n",
              "      <td>824.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>10</td>\n",
              "      <td>5</td>\n",
              "      <td>bogo</td>\n",
              "      <td>10</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>9b98b8c7a33c4b65b9aebfe6a799e6d9</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>60000.0</td>\n",
              "      <td>771.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5</td>\n",
              "      <td>7</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>60000.0</td>\n",
              "      <td>771.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5</td>\n",
              "      <td>5</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>004b041fbfe44859945daa2c7f79ee64</td>\n",
              "      <td>f19421c1d4aa40978ebb69ca19b0e20d</td>\n",
              "      <td>1</td>\n",
              "      <td>55</td>\n",
              "      <td>74000.0</td>\n",
              "      <td>158.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5</td>\n",
              "      <td>5</td>\n",
              "      <td>bogo</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             person                          offer_id  \\\n",
              "0  0011e0d4e6b944f998e987f904e8c1e5  9b98b8c7a33c4b65b9aebfe6a799e6d9   \n",
              "1  0020c2b971eb4e9188eac86d93036a77  4d5c57ea9a6940dd891ad53e9dbe8da0   \n",
              "2  0020ccbbb6d84e358d3414a3ff76cffd  9b98b8c7a33c4b65b9aebfe6a799e6d9   \n",
              "3  0020ccbbb6d84e358d3414a3ff76cffd  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "4  004b041fbfe44859945daa2c7f79ee64  f19421c1d4aa40978ebb69ca19b0e20d   \n",
              "\n",
              "   offer_responded  age   income  membership_days  offer_received_cnt  \\\n",
              "0                1   40  57000.0             39.0                   1   \n",
              "1                1   59  90000.0            824.0                   1   \n",
              "2                1   24  60000.0            771.0                   1   \n",
              "3                1   24  60000.0            771.0                   1   \n",
              "4                1   55  74000.0            158.0                   1   \n",
              "\n",
              "   amount_invalid  difficulty  duration offer_type  reward  web  email  \\\n",
              "0             0.0           5         7       bogo       5    1      1   \n",
              "1             0.0          10         5       bogo      10    1      1   \n",
              "2             0.0           5         7       bogo       5    1      1   \n",
              "3             0.0           5         5       bogo       5    1      1   \n",
              "4             0.0           5         5       bogo       5    1      1   \n",
              "\n",
              "   mobile  social  gender_F  gender_M  gender_O  \n",
              "0       1       0         0         0         1  \n",
              "1       1       1         1         0         0  \n",
              "2       1       0         1         0         0  \n",
              "3       1       1         1         0         0  \n",
              "4       1       1         1         0         0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L_M8FPoFyBKY",
        "colab_type": "code",
        "colab": {},
        "outputId": "86fb3833-a455-456d-f4fa-f5a91b4df183"
      },
      "source": [
        "# quick view on discount_offer data\n",
        "discount_offer.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>person</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>offer_responded</th>\n",
              "      <th>age</th>\n",
              "      <th>gender</th>\n",
              "      <th>income</th>\n",
              "      <th>membership_days</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0011e0d4e6b944f998e987f904e8c1e5</td>\n",
              "      <td>0b1e1539f2cc45b7b9fa7c272da2e1d7</td>\n",
              "      <td>1</td>\n",
              "      <td>40</td>\n",
              "      <td>O</td>\n",
              "      <td>57000.0</td>\n",
              "      <td>39.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0011e0d4e6b944f998e987f904e8c1e5</td>\n",
              "      <td>2298d6c36e964ae4a3e7e9706d1fb8c2</td>\n",
              "      <td>1</td>\n",
              "      <td>40</td>\n",
              "      <td>O</td>\n",
              "      <td>57000.0</td>\n",
              "      <td>39.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0020c2b971eb4e9188eac86d93036a77</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>1</td>\n",
              "      <td>59</td>\n",
              "      <td>F</td>\n",
              "      <td>90000.0</td>\n",
              "      <td>824.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>2298d6c36e964ae4a3e7e9706d1fb8c2</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>F</td>\n",
              "      <td>60000.0</td>\n",
              "      <td>771.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>003d66b6608740288d6cc97a6903f4f0</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>1</td>\n",
              "      <td>26</td>\n",
              "      <td>F</td>\n",
              "      <td>73000.0</td>\n",
              "      <td>566.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             person                          offer_id  \\\n",
              "0  0011e0d4e6b944f998e987f904e8c1e5  0b1e1539f2cc45b7b9fa7c272da2e1d7   \n",
              "1  0011e0d4e6b944f998e987f904e8c1e5  2298d6c36e964ae4a3e7e9706d1fb8c2   \n",
              "2  0020c2b971eb4e9188eac86d93036a77  fafdcd668e3743c1bb461111dcafc2a4   \n",
              "3  0020ccbbb6d84e358d3414a3ff76cffd  2298d6c36e964ae4a3e7e9706d1fb8c2   \n",
              "4  003d66b6608740288d6cc97a6903f4f0  fafdcd668e3743c1bb461111dcafc2a4   \n",
              "\n",
              "   offer_responded  age gender   income  membership_days  \n",
              "0                1   40      O  57000.0             39.0  \n",
              "1                1   40      O  57000.0             39.0  \n",
              "2                1   59      F  90000.0            824.0  \n",
              "3                1   24      F  60000.0            771.0  \n",
              "4                1   26      F  73000.0            566.0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2-kopZFSyBKZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# do the same processing on discount data\n",
        "# merge to get offers received count and invalid amount transacted \n",
        "discount_offer = discount_offer.merge(offer_cnt[['person','offer_id','offer_received_cnt']],how='left',on=['person','offer_id'])\n",
        "discount_offer = discount_offer.merge(transactions_not_related[['person','offer_id','amount_invalid']],how='left',on=['person','offer_id'])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z9CWn1JHyBKb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# fill missing values for amount_invalid with 0\n",
        "discount_offer['amount_invalid']=discount_offer['amount_invalid'].fillna(value=0)\n",
        "discount_offer.dropna(inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oTCFFEY-yBKc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# merge with portfolio to get offer details\n",
        "discount_offer= discount_offer.merge(portfolio,how='left',on='offer_id')\n",
        "\n",
        "# convert channels into categorical variables\n",
        "channels = discount_offer['channels'].apply(pd.Series)\n",
        "channels = channels.rename(columns={0:'web',1:'email',2:'mobile',3:'social'})\n",
        "discount_offer = pd.concat([discount_offer[:], channels[:]], axis=1)\n",
        "rename('web',discount_offer)\n",
        "rename('email',discount_offer)\n",
        "rename('mobile',discount_offer)\n",
        "rename('social',discount_offer)\n",
        "discount_offer = discount_offer.drop(['channels'], axis=1, inplace=False)\n",
        "\n",
        "#convert gender into categorical variables\n",
        "discount_offer=dummy(discount_offer,'gender')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ziWgEp_SyBKd",
        "colab_type": "code",
        "colab": {},
        "outputId": "b841185d-5de4-4a80-cea4-ad8fdacd681c"
      },
      "source": [
        "# quick view on discount_offer data\n",
        "discount_offer.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>person</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>offer_responded</th>\n",
              "      <th>age</th>\n",
              "      <th>income</th>\n",
              "      <th>membership_days</th>\n",
              "      <th>offer_received_cnt</th>\n",
              "      <th>amount_invalid</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>duration</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward</th>\n",
              "      <th>web</th>\n",
              "      <th>email</th>\n",
              "      <th>mobile</th>\n",
              "      <th>social</th>\n",
              "      <th>gender_F</th>\n",
              "      <th>gender_M</th>\n",
              "      <th>gender_O</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0011e0d4e6b944f998e987f904e8c1e5</td>\n",
              "      <td>0b1e1539f2cc45b7b9fa7c272da2e1d7</td>\n",
              "      <td>1</td>\n",
              "      <td>40</td>\n",
              "      <td>57000.0</td>\n",
              "      <td>39.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>20</td>\n",
              "      <td>10</td>\n",
              "      <td>discount</td>\n",
              "      <td>5</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0011e0d4e6b944f998e987f904e8c1e5</td>\n",
              "      <td>2298d6c36e964ae4a3e7e9706d1fb8c2</td>\n",
              "      <td>1</td>\n",
              "      <td>40</td>\n",
              "      <td>57000.0</td>\n",
              "      <td>39.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7</td>\n",
              "      <td>7</td>\n",
              "      <td>discount</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0020c2b971eb4e9188eac86d93036a77</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>1</td>\n",
              "      <td>59</td>\n",
              "      <td>90000.0</td>\n",
              "      <td>824.0</td>\n",
              "      <td>2</td>\n",
              "      <td>0.0</td>\n",
              "      <td>10</td>\n",
              "      <td>10</td>\n",
              "      <td>discount</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>2298d6c36e964ae4a3e7e9706d1fb8c2</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>60000.0</td>\n",
              "      <td>771.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7</td>\n",
              "      <td>7</td>\n",
              "      <td>discount</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>003d66b6608740288d6cc97a6903f4f0</td>\n",
              "      <td>fafdcd668e3743c1bb461111dcafc2a4</td>\n",
              "      <td>1</td>\n",
              "      <td>26</td>\n",
              "      <td>73000.0</td>\n",
              "      <td>566.0</td>\n",
              "      <td>2</td>\n",
              "      <td>0.0</td>\n",
              "      <td>10</td>\n",
              "      <td>10</td>\n",
              "      <td>discount</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             person                          offer_id  \\\n",
              "0  0011e0d4e6b944f998e987f904e8c1e5  0b1e1539f2cc45b7b9fa7c272da2e1d7   \n",
              "1  0011e0d4e6b944f998e987f904e8c1e5  2298d6c36e964ae4a3e7e9706d1fb8c2   \n",
              "2  0020c2b971eb4e9188eac86d93036a77  fafdcd668e3743c1bb461111dcafc2a4   \n",
              "3  0020ccbbb6d84e358d3414a3ff76cffd  2298d6c36e964ae4a3e7e9706d1fb8c2   \n",
              "4  003d66b6608740288d6cc97a6903f4f0  fafdcd668e3743c1bb461111dcafc2a4   \n",
              "\n",
              "   offer_responded  age   income  membership_days  offer_received_cnt  \\\n",
              "0                1   40  57000.0             39.0                   1   \n",
              "1                1   40  57000.0             39.0                   1   \n",
              "2                1   59  90000.0            824.0                   2   \n",
              "3                1   24  60000.0            771.0                   1   \n",
              "4                1   26  73000.0            566.0                   2   \n",
              "\n",
              "   amount_invalid  difficulty  duration offer_type  reward  web  email  \\\n",
              "0             0.0          20        10   discount       5    1      1   \n",
              "1             0.0           7         7   discount       3    1      1   \n",
              "2             0.0          10        10   discount       2    1      1   \n",
              "3             0.0           7         7   discount       3    1      1   \n",
              "4             0.0          10        10   discount       2    1      1   \n",
              "\n",
              "   mobile  social  gender_F  gender_M  gender_O  \n",
              "0       0       0         0         0         1  \n",
              "1       1       1         0         0         1  \n",
              "2       1       1         1         0         0  \n",
              "3       1       1         1         0         0  \n",
              "4       1       1         1         0         0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 66
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HlfADOmJyBKj",
        "colab_type": "code",
        "colab": {},
        "outputId": "ca6d9149-ec86-4a24-b74e-c368d06e8a23"
      },
      "source": [
        "# quick view on informational_offer data\n",
        "informational_offer.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>person</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>offer_responded</th>\n",
              "      <th>age</th>\n",
              "      <th>gender</th>\n",
              "      <th>income</th>\n",
              "      <th>membership_days</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>1</td>\n",
              "      <td>33</td>\n",
              "      <td>M</td>\n",
              "      <td>72000.0</td>\n",
              "      <td>506.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>1</td>\n",
              "      <td>33</td>\n",
              "      <td>M</td>\n",
              "      <td>72000.0</td>\n",
              "      <td>506.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0020c2b971eb4e9188eac86d93036a77</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>1</td>\n",
              "      <td>59</td>\n",
              "      <td>F</td>\n",
              "      <td>90000.0</td>\n",
              "      <td>824.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>F</td>\n",
              "      <td>60000.0</td>\n",
              "      <td>771.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>003d66b6608740288d6cc97a6903f4f0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>1</td>\n",
              "      <td>26</td>\n",
              "      <td>F</td>\n",
              "      <td>73000.0</td>\n",
              "      <td>566.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             person                          offer_id  \\\n",
              "0  0009655768c64bdeb2e877511632db8f  3f207df678b143eea3cee63160fa8bed   \n",
              "1  0009655768c64bdeb2e877511632db8f  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2  0020c2b971eb4e9188eac86d93036a77  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3  0020ccbbb6d84e358d3414a3ff76cffd  5a8bc65990b245e5a138643cd4eb9837   \n",
              "4  003d66b6608740288d6cc97a6903f4f0  3f207df678b143eea3cee63160fa8bed   \n",
              "\n",
              "   offer_responded  age gender   income  membership_days  \n",
              "0                1   33      M  72000.0            506.0  \n",
              "1                1   33      M  72000.0            506.0  \n",
              "2                1   59      F  90000.0            824.0  \n",
              "3                1   24      F  60000.0            771.0  \n",
              "4                1   26      F  73000.0            566.0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 67
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aOxmVVc4yBKn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# processing on informational data\n",
        "# merge with portfolio to get offer details\n",
        "informational_offer = informational_offer.merge(portfolio,how='left',on='offer_id')\n",
        "\n",
        "# reset index for informational_offer\n",
        "informational_offer = informational_offer.reset_index().drop(['index'],axis=1,inplace=False)\n",
        "\n",
        "# split the channel column\n",
        "informational_offer['email']= np.nan\n",
        "informational_offer.loc[pd.Series(['email' in informational_offer['channels'][x] for x in range(len(informational_offer['channels']))]),'email']=1\n",
        "informational_offer['email']=informational_offer['email'].fillna(value=0)\n",
        "\n",
        "informational_offer['mobile']= np.nan\n",
        "informational_offer.loc[pd.Series(['mobile' in informational_offer['channels'][x] for x in range(len(informational_offer['channels']))]),'mobile']=1\n",
        "informational_offer['mobile']=informational_offer['mobile'].fillna(value=0)\n",
        "\n",
        "informational_offer['social']= np.nan\n",
        "informational_offer.loc[pd.Series(['social' in informational_offer['channels'][x] for x in range(len(informational_offer['channels']))]),'social']=1\n",
        "informational_offer['social']=informational_offer['social'].fillna(value=0)\n",
        "\n",
        "# drop the useless columns\n",
        "informational_offer.drop(['channels'], axis=1, inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fhZFjZdLyBKq",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# merge the informational offer dataframe\n",
        "informational_offer = informational_offer.merge(offer_cnt[['person','offer_id','offer_received_cnt']],how='left',on=['person','offer_id'])\n",
        "informational_offer = informational_offer.merge(transactions_not_related[['person','offer_id','amount_invalid']],how='left',on=['person','offer_id'])\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T2_nVlMoyBKr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# fill missing values for amount_invalid and offer_received_cnt with 0\n",
        "informational_offer['amount_invalid']=informational_offer['amount_invalid'].fillna(value=0)\n",
        "\n",
        "#drop income and gender null rows\n",
        "informational_offer.dropna(inplace=True);    \n",
        "\n",
        "#convert gender column into dummy variables\n",
        "informational_offer=dummy(informational_offer,'gender')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SjVoyyb_yBKt",
        "colab_type": "code",
        "colab": {},
        "outputId": "2aa588da-9274-404e-8663-5d11dfe2b904"
      },
      "source": [
        "# quick view on informational_offer data\n",
        "informational_offer.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>person</th>\n",
              "      <th>offer_id</th>\n",
              "      <th>offer_responded</th>\n",
              "      <th>age</th>\n",
              "      <th>income</th>\n",
              "      <th>membership_days</th>\n",
              "      <th>difficulty</th>\n",
              "      <th>duration</th>\n",
              "      <th>offer_type</th>\n",
              "      <th>reward</th>\n",
              "      <th>email</th>\n",
              "      <th>mobile</th>\n",
              "      <th>social</th>\n",
              "      <th>offer_received_cnt</th>\n",
              "      <th>amount_invalid</th>\n",
              "      <th>gender_F</th>\n",
              "      <th>gender_M</th>\n",
              "      <th>gender_O</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>1</td>\n",
              "      <td>33</td>\n",
              "      <td>72000.0</td>\n",
              "      <td>506.0</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0009655768c64bdeb2e877511632db8f</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>1</td>\n",
              "      <td>33</td>\n",
              "      <td>72000.0</td>\n",
              "      <td>506.0</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0020c2b971eb4e9188eac86d93036a77</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>1</td>\n",
              "      <td>59</td>\n",
              "      <td>90000.0</td>\n",
              "      <td>824.0</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.00</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0020ccbbb6d84e358d3414a3ff76cffd</td>\n",
              "      <td>5a8bc65990b245e5a138643cd4eb9837</td>\n",
              "      <td>1</td>\n",
              "      <td>24</td>\n",
              "      <td>60000.0</td>\n",
              "      <td>771.0</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>16.27</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>003d66b6608740288d6cc97a6903f4f0</td>\n",
              "      <td>3f207df678b143eea3cee63160fa8bed</td>\n",
              "      <td>1</td>\n",
              "      <td>26</td>\n",
              "      <td>73000.0</td>\n",
              "      <td>566.0</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>informational</td>\n",
              "      <td>0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.00</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             person                          offer_id  \\\n",
              "0  0009655768c64bdeb2e877511632db8f  3f207df678b143eea3cee63160fa8bed   \n",
              "1  0009655768c64bdeb2e877511632db8f  5a8bc65990b245e5a138643cd4eb9837   \n",
              "2  0020c2b971eb4e9188eac86d93036a77  5a8bc65990b245e5a138643cd4eb9837   \n",
              "3  0020ccbbb6d84e358d3414a3ff76cffd  5a8bc65990b245e5a138643cd4eb9837   \n",
              "4  003d66b6608740288d6cc97a6903f4f0  3f207df678b143eea3cee63160fa8bed   \n",
              "\n",
              "   offer_responded  age   income  membership_days  difficulty  duration  \\\n",
              "0                1   33  72000.0            506.0           0         4   \n",
              "1                1   33  72000.0            506.0           0         3   \n",
              "2                1   59  90000.0            824.0           0         3   \n",
              "3                1   24  60000.0            771.0           0         3   \n",
              "4                1   26  73000.0            566.0           0         4   \n",
              "\n",
              "      offer_type  reward  email  mobile  social  offer_received_cnt  \\\n",
              "0  informational       0    1.0     1.0     0.0                 1.0   \n",
              "1  informational       0    1.0     1.0     1.0                 1.0   \n",
              "2  informational       0    1.0     1.0     1.0                 1.0   \n",
              "3  informational       0    1.0     1.0     1.0                 1.0   \n",
              "4  informational       0    1.0     1.0     0.0                 1.0   \n",
              "\n",
              "   amount_invalid  gender_F  gender_M  gender_O  \n",
              "0            0.00         0         1         0  \n",
              "1            0.00         0         1         0  \n",
              "2            0.00         1         0         0  \n",
              "3           16.27         1         0         0  \n",
              "4            0.00         1         0         0  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 71
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iM6yFm7HyBKv",
        "colab_type": "text"
      },
      "source": [
        "## Building model\n",
        "After pre-processing the data, the next step we'll start to implement models to figure out which factors affect most whether the customer will respond to the offer or not. And this project also attempts to predict whether the customer will respond to the different types of offers or not.\n",
        "\n",
        "Therefore, we'll use the 'offer_responded' flag in the dataset to build models to predict if the customer will respond to the offer of not. Here we will choose the basic tree model as a baseline which will help explain the feature importance better so that we can get some insight into what factors affect customer's behavior most. Meanwhile, I'll also choose random forest as an alternate model to compare the baseline model is as an improvement over simple ensemble bagging of decision trees, in order to drive towards a high accuracy in training the model."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p5uSh3tOyBKv",
        "colab_type": "text"
      },
      "source": [
        "### Model implementation preparation\n",
        "- Prepare the dateset, set the features variable and target columns"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Uc4Q9A_7yBKv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def data_prep(df,drop_cols_prep):\n",
        "    '''\n",
        "    inputs:\n",
        "    - df: prepared dataframe for modeling \n",
        "    \n",
        "    outputs:\n",
        "    - Returns 2 dataframes - features and target dataframes\n",
        "    '''\n",
        "    # Split the data into features and target label\n",
        "    target = df['offer_responded']\n",
        "    features = df.drop(drop_cols_prep, axis=1, inplace=False)\n",
        "    return features,target"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FY5b_sqdyBKx",
        "colab_type": "text"
      },
      "source": [
        "- Split the data into training and test sets"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cUIg5qmSyBKx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def model_pipeline(features,target):\n",
        "    '''\n",
        "    inputs:\n",
        "    - features & target dataframe \n",
        "    \n",
        "    outputs:\n",
        "    - Splits features and target dataframe to train and test sets, performs feature scaling on both datasets.\n",
        "    - Outputs X_train, X_test, y_train and y_test dataframes\n",
        "    '''\n",
        "    \n",
        "    #split into training and test sets\n",
        "    X_train, X_test, y_train, y_test = train_test_split(features,target, test_size=0.20, random_state=42)\n",
        "\n",
        "    #fit and transform scaling on training data\n",
        "    scaler=StandardScaler()\n",
        "    X_train=scaler.fit_transform(X_train)\n",
        "\n",
        "    #scale test data\n",
        "    X_test=scaler.transform(X_test)\n",
        "    return X_train,X_test,y_train, y_test"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rD05uYIqyBKz",
        "colab_type": "text"
      },
      "source": [
        "- Create function to execute the model for different offer types"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qKEeViNQyBKz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# reference: Udacity -- 'Finding Donors for Charity ML' project\n",
        "# reference: Udacity -- 'Creating Customer Segments with Arvato' project\n",
        "def train_predict(model, X_train, y_train, X_test, y_test): \n",
        "    '''\n",
        "    inputs:\n",
        "       - model: the model to be trained and predicted on\n",
        "       - sample_size: the size of samples (number) to be drawn from training set\n",
        "       - X_train: features training set\n",
        "       - y_train: review_scores_rating training set\n",
        "       - X_test: features testing set\n",
        "       - y_test: review_scores_rating testing set\n",
        "    '''\n",
        "    results = {}\n",
        "    \n",
        "    #Fit the model to the training data and get training time\n",
        "    start = time() \n",
        "    model = model.fit(X_train, y_train)\n",
        "    end = time() \n",
        "    results['train_time'] = end-start\n",
        "    \n",
        "    # Get predictions on the test set(X_test), then get predictions on first 300 training samples\n",
        "    start = time() \n",
        "    predictions_test = model.predict(X_test)\n",
        "    predictions_train = model.predict(X_train)\n",
        "    end = time() \n",
        "    \n",
        "    # Calculate the total prediction time\n",
        "    results['pred_time'] = end-start\n",
        "    \n",
        "    #add training accuracy to results\n",
        "    results['training_score']=model.score(X_train,y_train)\n",
        "    \n",
        "    #add testing accuracy to results\n",
        "    results['testing_score']=model.score(X_test,y_test)\n",
        "     \n",
        "    print(\"{} trained on {} samples.\".format(model.__class__.__name__, len(y_train)))\n",
        "    print(\"MSE_train: %.4f\" % mean_squared_error(y_train,predictions_train))\n",
        "    print(\"MSE_test: %.4f\" % mean_squared_error(y_test,predictions_test))\n",
        "    print(\"Training accuracy:%.4f\" % results['training_score'])\n",
        "    print(\"Test accuracy:%.4f\" % results['testing_score'])\n",
        "    print(classification_report(y_test, predictions_test,digits=4))\n",
        "    return results"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KVmbb1mdyBK0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def run_model(clf1,clf2,name):\n",
        "    '''\n",
        "    inputs:\n",
        "    - clf1: first classifier model\n",
        "    - clf2: 2nd classifier model for comparison\n",
        "    - name: name of models for comparison\n",
        "    \n",
        "    outputs:\n",
        "    - Dataframe of results from model training and prediction\n",
        "    '''\n",
        "    \n",
        "    # Collect results from models\n",
        "    results = {}\n",
        "    for clf in [clf1, clf2]:\n",
        "        clf_name = clf.__class__.__name__ + '_' +name\n",
        "        results[clf_name] = {}\n",
        "        results[clf_name]= train_predict(clf, X_train, y_train, X_test, y_test)\n",
        "    return pd.DataFrame(results)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kQXGabE-yBK3",
        "colab_type": "text"
      },
      "source": [
        "### Performance Metrics\n",
        "Since the project is building classification model, I choose both accuracy and F1 score as the model evaluation metric. The reason why I choose both metrics is sometimes when the dataset is imbalanced, the accuracy only couldn't objectively show how the model is performing on the dataset, while F1 score provides a better sense of model performance compared to purely accuracy as takes both false positives and false negatives in the calculation. With an imbalanced class distribution, F1 may be more useful than accuracy.\n",
        "\n",
        "Also, since the F1 score is based on the harmonic mean of precision and recall, and focuses on positive cases. For the Starbucks app here, it would be fine as we would prioritise more on whether offers are effective, and less focus on why offers are ineffective."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yTI16XrayBK3",
        "colab_type": "text"
      },
      "source": [
        "### Initial the model baseline\n",
        "At this point, we will firstly use default parameters for the baseline model and will tune the parameters in the later tuning steps if needed."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "R8S7pkAhyBK3",
        "colab_type": "text"
      },
      "source": [
        "- __BOGO model__"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k9eGnm4AyBK4",
        "colab_type": "code",
        "colab": {},
        "outputId": "e8a10d34-f6fc-454a-e768-967f209641b1"
      },
      "source": [
        "# implement the model for BOGO offer\n",
        "drop_cols_prep=['person','offer_id','offer_responded','offer_type']\n",
        "features,target=data_prep(bogo_offer,drop_cols_prep)\n",
        "X_train, X_test, y_train, y_test=model_pipeline(features,target)\n",
        "\n",
        "# initialize the model - baseline is DT model, bogo_1 model is RF model\n",
        "baseline = DecisionTreeClassifier(criterion='entropy',max_depth=5,random_state=2,min_samples_split=90,min_samples_leaf=50)\n",
        "bogo_1 = RandomForestClassifier(random_state=2,max_depth= 11, max_features= 'auto',min_samples_split= 10,n_estimators=20,min_samples_leaf=20)\n",
        "\n",
        "results=run_model(baseline,bogo_1,'bogo_1')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "DecisionTreeClassifier trained on 9829 samples.\n",
            "MSE_train: 0.1770\n",
            "MSE_test: 0.1823\n",
            "Training accuracy:0.8230\n",
            "Test accuracy:0.8177\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     0.4797    0.2694    0.3450       438\n",
            "          1     0.8553    0.9366    0.8941      2020\n",
            "\n",
            "avg / total     0.7884    0.8177    0.7963      2458\n",
            "\n",
            "RandomForestClassifier trained on 9829 samples.\n",
            "MSE_train: 0.1670\n",
            "MSE_test: 0.1786\n",
            "Training accuracy:0.8330\n",
            "Test accuracy:0.8214\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     0.4906    0.0594    0.1059       438\n",
            "          1     0.8287    0.9866    0.9008      2020\n",
            "\n",
            "avg / total     0.7684    0.8214    0.7591      2458\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H6Pu-g1nyBK7",
        "colab_type": "text"
      },
      "source": [
        "As shown above, the accuracy of both models is good (more than 80%) for initial model implementation. But the F1 score is a bit lower than accuracy which may be tuned better in the later steps. From the result we can see that Decision Tree's F1 performs a little better than Random Forest, which means that decision tree is doing slightly better compared to random forest at not misclassifying negative events as positive (meaning, misclassifying people on which offers are ineffective, as people on which offers would be effective).\n",
        "\n",
        "But the overall higher accuracy compared to F1 score indicates that the model is predicting the positive case (i.e. where an offer is responded) more accurately compared to predicting the negative cases (i.e. where an offer is not reponded), which is expected given the uneven classes.\n",
        "\n",
        "However, in this project context, there's not big hurt to send out some more offers to people who are not gonna respond in the end. Therefore, here can still select the random forest with slightly better accuracy right now. So, right now, will stay on random forest model."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GjS6-I3myBK7",
        "colab_type": "text"
      },
      "source": [
        "- __Discount Offer model__"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gy0VD30OyBK7",
        "colab_type": "code",
        "colab": {},
        "outputId": "5f33aa1d-ac71-4832-d58c-f9ef3afaad9a"
      },
      "source": [
        "# instantiate the model for discount offer\n",
        "drop_cols_prep=['person','offer_id','offer_responded','offer_type']\n",
        "features,target=data_prep(discount_offer,drop_cols_prep)\n",
        "X_train, X_test, y_train, y_test=model_pipeline(features,target)\n",
        "\n",
        "#Initialize the model\n",
        "discount_1 = RandomForestClassifier(random_state=2,max_depth= 20, max_features= 'auto',min_samples_split= 10,n_estimators=20,min_samples_leaf=10)\n",
        "results=pd.concat([results[:],run_model(baseline,discount_1,'discount_1')],axis=1)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "DecisionTreeClassifier trained on 10179 samples.\n",
            "MSE_train: 0.1371\n",
            "MSE_test: 0.1277\n",
            "Training accuracy:0.8629\n",
            "Test accuracy:0.8723\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     0.0000    0.0000    0.0000       325\n",
            "          1     0.8723    1.0000    0.9318      2220\n",
            "\n",
            "avg / total     0.7609    0.8723    0.8128      2545\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/opt/conda/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
            "  'precision', 'predicted', average, warn_for)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "RandomForestClassifier trained on 10179 samples.\n",
            "MSE_train: 0.1313\n",
            "MSE_test: 0.1277\n",
            "Training accuracy:0.8687\n",
            "Test accuracy:0.8723\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     0.5000    0.0062    0.0122       325\n",
            "          1     0.8729    0.9991    0.9317      2220\n",
            "\n",
            "avg / total     0.8253    0.8723    0.8143      2545\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "m7pNIrydyBK9",
        "colab_type": "text"
      },
      "source": [
        "As shown above, the random forest performs quiet similar as random forest, but decision tree got a slightly lower F1 score. The F1 score for these models are lower overall compared to the Accuracy score. \n",
        "\n",
        "As mentioned in previous steps, it's more important in this business case for the model predicting positive cases accurately, so would rather go with a higher accuracy model where F1 score is higher, for which random forest classifier has better performance (0.814 vs 0.812)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MZpumOcMyBK9",
        "colab_type": "text"
      },
      "source": [
        "- __Informational offer model__"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HV8CpCrAyBK9",
        "colab_type": "code",
        "colab": {},
        "outputId": "bdf87fae-c6c1-4fad-f96e-78ca8c982ede"
      },
      "source": [
        "# implement model for informational offer\n",
        "features,target=data_prep(informational_offer,drop_cols_prep)\n",
        "X_train, X_test, y_train, y_test=model_pipeline(features,target)\n",
        "\n",
        "#Initialize the model\n",
        "info_1 = RandomForestClassifier(random_state=5,criterion='gini',max_depth= 20, max_features= 'auto',min_samples_split= 10,n_estimators=20,min_samples_leaf=10)\n",
        "results=pd.concat([results[:],run_model(baseline,info_1,'info_1')],axis=1)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "DecisionTreeClassifier trained on 5585 samples.\n",
            "MSE_train: 0.2462\n",
            "MSE_test: 0.2541\n",
            "Training accuracy:0.7538\n",
            "Test accuracy:0.7459\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     0.5000    0.1127    0.1839       355\n",
            "          1     0.7608    0.9616    0.8495      1042\n",
            "\n",
            "avg / total     0.6945    0.7459    0.6804      1397\n",
            "\n",
            "RandomForestClassifier trained on 5585 samples.\n",
            "MSE_train: 0.2319\n",
            "MSE_test: 0.2520\n",
            "Training accuracy:0.7681\n",
            "Test accuracy:0.7480\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     0.5200    0.1099    0.1814       355\n",
            "          1     0.7610    0.9655    0.8511      1042\n",
            "\n",
            "avg / total     0.6997    0.7480    0.6809      1397\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fbveAp_xyBK_",
        "colab_type": "text"
      },
      "source": [
        "### Model tuning\n",
        "This section will attempt to tune the parameters of initial model using GridSearch method to get higher performance."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WpU1RycByBK_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# create the function to find best model results for each offer type\n",
        "def best_model(offer_type):\n",
        "    '''\n",
        "    input:\n",
        "    - offer_type: string of offer type name\n",
        "    output:\n",
        "    - dataframe containing results of best model so far\n",
        "    \n",
        "    '''\n",
        "    print(offer_type + ' RF model:')\n",
        "    return results.transpose()[results.transpose()['testing_score']==results.transpose()[results.transpose().index.str.contains(\"RandomForestClassifier_\"+offer_type)]['testing_score'].max()]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "efudCrNGyBLA",
        "colab_type": "text"
      },
      "source": [
        "In the tuning section, we will first use GridSearch to search for parameters that are likely to get better model performance."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M8n3-bY_yBLB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#define Grid Search function\n",
        "def rand_forest_param_selection(X,y):\n",
        "    '''\n",
        "    input:\n",
        "    - X,y: training datasets for X and y\n",
        "    output:\n",
        "    - dictionary with best parameters for random forest model\n",
        "    '''\n",
        "    \n",
        "    param_grid={'max_features': ['auto', 'sqrt'],\n",
        "                'max_depth' : [10,15],\n",
        "                'n_estimators': [10,20,25,30],\n",
        "                'min_samples_split': [10, 20],\n",
        "                'min_samples_leaf': [10,15],\n",
        "                }\n",
        "    grid_search = GridSearchCV(RandomForestClassifier(random_state=2), param_grid)\n",
        "    grid_search.fit(X, y)\n",
        "    grid_search.best_params_\n",
        "    return grid_search.best_params_"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Dd5WgolHyBLC",
        "colab_type": "code",
        "colab": {},
        "outputId": "d280190a-f859-4f6d-d24e-e0633d9b173d"
      },
      "source": [
        "#define BOGO dataset\n",
        "features,target=data_prep(bogo_offer,drop_cols_prep)\n",
        "X_train, X_test, y_train, y_test=model_pipeline(features,target)\n",
        "\n",
        "#run Grid Search\n",
        "rand_forest_param_selection(X_train, y_train)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'max_depth': 10,\n",
              " 'max_features': 'auto',\n",
              " 'min_samples_leaf': 10,\n",
              " 'min_samples_split': 10,\n",
              " 'n_estimators': 20}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 81
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9KAOCkzcyBLE",
        "colab_type": "code",
        "colab": {},
        "outputId": "535b98fc-4164-47de-cf12-f468ad449510"
      },
      "source": [
        "# use optimized parameters to rerun the model in previous step\n",
        "# initialize the model\n",
        "bogo_2 = RandomForestClassifier(random_state=2,max_depth= 10, max_features= 'auto',min_samples_split= 10,n_estimators=20,min_samples_leaf=10)\n",
        "\n",
        "results=pd.concat([results[:],run_model(baseline,bogo_2,'bogo_2')],axis=1)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "DecisionTreeClassifier trained on 9829 samples.\n",
            "MSE_train: 0.1770\n",
            "MSE_test: 0.1823\n",
            "Training accuracy:0.8230\n",
            "Test accuracy:0.8177\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     0.4797    0.2694    0.3450       438\n",
            "          1     0.8553    0.9366    0.8941      2020\n",
            "\n",
            "avg / total     0.7884    0.8177    0.7963      2458\n",
            "\n",
            "RandomForestClassifier trained on 9829 samples.\n",
            "MSE_train: 0.1613\n",
            "MSE_test: 0.1717\n",
            "Training accuracy:0.8387\n",
            "Test accuracy:0.8283\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     0.5870    0.1233    0.2038       438\n",
            "          1     0.8377    0.9812    0.9038      2020\n",
            "\n",
            "avg / total     0.7930    0.8283    0.7790      2458\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V1KrDHM1yBLG",
        "colab_type": "code",
        "colab": {},
        "outputId": "f982e538-82f2-4072-aa1a-4c07dccddccc"
      },
      "source": [
        "results[['RandomForestClassifier_bogo_1','RandomForestClassifier_bogo_2']]"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>RandomForestClassifier_bogo_1</th>\n",
              "      <th>RandomForestClassifier_bogo_2</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>pred_time</th>\n",
              "      <td>0.029760</td>\n",
              "      <td>0.030382</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>testing_score</th>\n",
              "      <td>0.821400</td>\n",
              "      <td>0.828316</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>train_time</th>\n",
              "      <td>0.146926</td>\n",
              "      <td>0.150937</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>training_score</th>\n",
              "      <td>0.833045</td>\n",
              "      <td>0.838742</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                RandomForestClassifier_bogo_1  RandomForestClassifier_bogo_2\n",
              "pred_time                            0.029760                       0.030382\n",
              "testing_score                        0.821400                       0.828316\n",
              "train_time                           0.146926                       0.150937\n",
              "training_score                       0.833045                       0.838742"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 83
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7Z-XWq3byBLH",
        "colab_type": "code",
        "colab": {},
        "outputId": "e3b56436-cdad-4f3e-983b-35caf3cbcb32"
      },
      "source": [
        "# best model for BOGO offer type\n",
        "best_model('bogo')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "bogo RF model:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>pred_time</th>\n",
              "      <th>testing_score</th>\n",
              "      <th>train_time</th>\n",
              "      <th>training_score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>RandomForestClassifier_bogo_2</th>\n",
              "      <td>0.030382</td>\n",
              "      <td>0.828316</td>\n",
              "      <td>0.150937</td>\n",
              "      <td>0.838742</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                               pred_time  testing_score  train_time  \\\n",
              "RandomForestClassifier_bogo_2   0.030382       0.828316    0.150937   \n",
              "\n",
              "                               training_score  \n",
              "RandomForestClassifier_bogo_2        0.838742  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 84
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "37jgY_9YyBLJ",
        "colab_type": "text"
      },
      "source": [
        "As shown above in the comparison, after using tune parameters, the test accuracy slightly improved from 0.833 to 0.838 and the F1 score increase from 0.759 to 0.779."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YISLDvDIyBLJ",
        "colab_type": "code",
        "colab": {},
        "outputId": "53677e55-2ec9-4f3d-aede-7730cbb3c001"
      },
      "source": [
        "# do the same tuning and refit steps on discount offer\n",
        "features,target=data_prep(discount_offer,drop_cols_prep)\n",
        "X_train, X_test, y_train, y_test=model_pipeline(features,target)\n",
        "\n",
        "# run Grid Search\n",
        "rand_forest_param_selection(X_train, y_train)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'max_depth': 10,\n",
              " 'max_features': 'auto',\n",
              " 'min_samples_leaf': 15,\n",
              " 'min_samples_split': 10,\n",
              " 'n_estimators': 30}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 85
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UbUnic11yBLK",
        "colab_type": "code",
        "colab": {},
        "outputId": "7687633e-dc9e-43ec-9340-78ae578d621d"
      },
      "source": [
        "# rerun the model with tuned parameters\n",
        "# initialize the model\n",
        "discount_2 = RandomForestClassifier(random_state=2,max_depth= 10, max_features= 'auto',min_samples_split= 10,n_estimators=30,min_samples_leaf=15)\n",
        "\n",
        "results=pd.concat([results[:],run_model(baseline,discount_2,'discount_2')],axis=1)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "DecisionTreeClassifier trained on 10179 samples.\n",
            "MSE_train: 0.1371\n",
            "MSE_test: 0.1277\n",
            "Training accuracy:0.8629\n",
            "Test accuracy:0.8723\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     0.0000    0.0000    0.0000       325\n",
            "          1     0.8723    1.0000    0.9318      2220\n",
            "\n",
            "avg / total     0.7609    0.8723    0.8128      2545\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/opt/conda/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
            "  'precision', 'predicted', average, warn_for)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "RandomForestClassifier trained on 10179 samples.\n",
            "MSE_train: 0.1350\n",
            "MSE_test: 0.1261\n",
            "Training accuracy:0.8650\n",
            "Test accuracy:0.8739\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     1.0000    0.0123    0.0243       325\n",
            "          1     0.8737    1.0000    0.9326      2220\n",
            "\n",
            "avg / total     0.8898    0.8739    0.8166      2545\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gUjN9E1CyBLL",
        "colab_type": "code",
        "colab": {},
        "outputId": "bdcb575d-be80-4732-9059-a359f0d051b0"
      },
      "source": [
        "results[['RandomForestClassifier_discount_1','RandomForestClassifier_discount_2']]"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>RandomForestClassifier_discount_1</th>\n",
              "      <th>RandomForestClassifier_discount_2</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>pred_time</th>\n",
              "      <td>0.035388</td>\n",
              "      <td>0.044937</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>testing_score</th>\n",
              "      <td>0.872299</td>\n",
              "      <td>0.873870</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>train_time</th>\n",
              "      <td>0.164510</td>\n",
              "      <td>0.231332</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>training_score</th>\n",
              "      <td>0.868749</td>\n",
              "      <td>0.865016</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                RandomForestClassifier_discount_1  \\\n",
              "pred_time                                0.035388   \n",
              "testing_score                            0.872299   \n",
              "train_time                               0.164510   \n",
              "training_score                           0.868749   \n",
              "\n",
              "                RandomForestClassifier_discount_2  \n",
              "pred_time                                0.044937  \n",
              "testing_score                            0.873870  \n",
              "train_time                               0.231332  \n",
              "training_score                           0.865016  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 87
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xuF8S-NzyBLM",
        "colab_type": "code",
        "colab": {},
        "outputId": "e4e4f452-3d73-43ea-b709-52f51291b52e"
      },
      "source": [
        "# best model for discount offer type\n",
        "best_model('discount')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "discount RF model:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>pred_time</th>\n",
              "      <th>testing_score</th>\n",
              "      <th>train_time</th>\n",
              "      <th>training_score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>RandomForestClassifier_discount_2</th>\n",
              "      <td>0.044937</td>\n",
              "      <td>0.87387</td>\n",
              "      <td>0.231332</td>\n",
              "      <td>0.865016</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                   pred_time  testing_score  train_time  \\\n",
              "RandomForestClassifier_discount_2   0.044937        0.87387    0.231332   \n",
              "\n",
              "                                   training_score  \n",
              "RandomForestClassifier_discount_2        0.865016  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 88
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lnIHzUFLyBLO",
        "colab_type": "text"
      },
      "source": [
        "As shown above in the comparison, after using tune parameters, the test accuracy slightly improved from 0.872 to 0.873 and the F1 score increase from 0.814 to 0.816."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XN_vqxkkyBLO",
        "colab_type": "code",
        "colab": {},
        "outputId": "54a74d26-fb01-4aa2-8a90-4ce793cd5a81"
      },
      "source": [
        "# model tuning for informational offer model\n",
        "features,target=data_prep(informational_offer,drop_cols_prep)\n",
        "X_train, X_test, y_train, y_test=model_pipeline(features,target)\n",
        "\n",
        "#run Grid Search\n",
        "rand_forest_param_selection(X_train, y_train)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'max_depth': 10,\n",
              " 'max_features': 'auto',\n",
              " 'min_samples_leaf': 10,\n",
              " 'min_samples_split': 10,\n",
              " 'n_estimators': 10}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 89
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_x5ooreHyBLR",
        "colab_type": "code",
        "colab": {},
        "outputId": "88825de8-6506-45dd-9763-7ce07d17ff61"
      },
      "source": [
        "# rerun the model with selected paramenters\n",
        "info_2 = RandomForestClassifier(random_state=2,max_depth= 10, max_features= 'auto',min_samples_split= 10,n_estimators=10,min_samples_leaf=10)\n",
        "results=pd.concat([results[:],run_model(baseline,info_2,'info_2')],axis=1)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "DecisionTreeClassifier trained on 5585 samples.\n",
            "MSE_train: 0.2462\n",
            "MSE_test: 0.2541\n",
            "Training accuracy:0.7538\n",
            "Test accuracy:0.7459\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     0.5000    0.1127    0.1839       355\n",
            "          1     0.7608    0.9616    0.8495      1042\n",
            "\n",
            "avg / total     0.6945    0.7459    0.6804      1397\n",
            "\n",
            "RandomForestClassifier trained on 5585 samples.\n",
            "MSE_train: 0.2378\n",
            "MSE_test: 0.2470\n",
            "Training accuracy:0.7622\n",
            "Test accuracy:0.7530\n",
            "             precision    recall  f1-score   support\n",
            "\n",
            "          0     0.5893    0.0930    0.1606       355\n",
            "          1     0.7599    0.9779    0.8552      1042\n",
            "\n",
            "avg / total     0.7165    0.7530    0.6787      1397\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "63EjNKu2yBLT",
        "colab_type": "code",
        "colab": {},
        "outputId": "e55433cc-edfe-4da3-d8bf-6767de6add67"
      },
      "source": [
        "results[['RandomForestClassifier_info_1','RandomForestClassifier_info_2']]"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>RandomForestClassifier_info_1</th>\n",
              "      <th>RandomForestClassifier_info_2</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>pred_time</th>\n",
              "      <td>0.019158</td>\n",
              "      <td>0.009385</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>testing_score</th>\n",
              "      <td>0.748031</td>\n",
              "      <td>0.753042</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>train_time</th>\n",
              "      <td>0.094609</td>\n",
              "      <td>0.045562</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>training_score</th>\n",
              "      <td>0.768129</td>\n",
              "      <td>0.762220</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                RandomForestClassifier_info_1  RandomForestClassifier_info_2\n",
              "pred_time                            0.019158                       0.009385\n",
              "testing_score                        0.748031                       0.753042\n",
              "train_time                           0.094609                       0.045562\n",
              "training_score                       0.768129                       0.762220"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 91
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xv33FPxFyBLU",
        "colab_type": "code",
        "colab": {},
        "outputId": "dafa3126-ffe7-44db-ddaf-4ee20caef0a7"
      },
      "source": [
        "# best model for informational offer type\n",
        "best_model('info')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "info RF model:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>pred_time</th>\n",
              "      <th>testing_score</th>\n",
              "      <th>train_time</th>\n",
              "      <th>training_score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>RandomForestClassifier_info_2</th>\n",
              "      <td>0.009385</td>\n",
              "      <td>0.753042</td>\n",
              "      <td>0.045562</td>\n",
              "      <td>0.76222</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                               pred_time  testing_score  train_time  \\\n",
              "RandomForestClassifier_info_2   0.009385       0.753042    0.045562   \n",
              "\n",
              "                               training_score  \n",
              "RandomForestClassifier_info_2         0.76222  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 92
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_TLh9_rDyBLW",
        "colab_type": "text"
      },
      "source": [
        "As shown above in the comparison, after using tune parameters, the test accuray slightly improved from 0.748 to 0.753 and the F1 score increase from 0.681 to 0.678."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e7-AktfLyBLX",
        "colab_type": "text"
      },
      "source": [
        "### View the feature importance\n",
        "Next, we'll look at the model's result and see if there's any insight of main factors which decide whether customers will respond to offers we could get by investigating feature importance."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bcKEvjtOyBLX",
        "colab_type": "code",
        "colab": {},
        "outputId": "695c67a3-ce25-4334-f787-cfc75be410e7"
      },
      "source": [
        "# models summary\n",
        "best_model('bogo').append([best_model('discount'),best_model('info')]).transpose()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "bogo RF model:\n",
            "discount RF model:\n",
            "info RF model:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>RandomForestClassifier_bogo_2</th>\n",
              "      <th>RandomForestClassifier_discount_2</th>\n",
              "      <th>RandomForestClassifier_info_2</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>pred_time</th>\n",
              "      <td>0.030382</td>\n",
              "      <td>0.044937</td>\n",
              "      <td>0.009385</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>testing_score</th>\n",
              "      <td>0.828316</td>\n",
              "      <td>0.873870</td>\n",
              "      <td>0.753042</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>train_time</th>\n",
              "      <td>0.150937</td>\n",
              "      <td>0.231332</td>\n",
              "      <td>0.045562</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>training_score</th>\n",
              "      <td>0.838742</td>\n",
              "      <td>0.865016</td>\n",
              "      <td>0.762220</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                RandomForestClassifier_bogo_2  \\\n",
              "pred_time                            0.030382   \n",
              "testing_score                        0.828316   \n",
              "train_time                           0.150937   \n",
              "training_score                       0.838742   \n",
              "\n",
              "                RandomForestClassifier_discount_2  \\\n",
              "pred_time                                0.044937   \n",
              "testing_score                            0.873870   \n",
              "train_time                               0.231332   \n",
              "training_score                           0.865016   \n",
              "\n",
              "                RandomForestClassifier_info_2  \n",
              "pred_time                            0.009385  \n",
              "testing_score                        0.753042  \n",
              "train_time                           0.045562  \n",
              "training_score                       0.762220  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 93
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": false,
        "id": "sNJvZOtGyBLZ",
        "colab_type": "code",
        "colab": {},
        "outputId": "5f22d37a-5aaa-49cb-edd7-9d6751bfb6e4"
      },
      "source": [
        "# display the feature importance\n",
        "# BOGO model\n",
        "drop_cols_prep=['person','offer_id','offer_responded','offer_type']\n",
        "features,target=data_prep(bogo_offer,drop_cols_prep)\n",
        "\n",
        "feature_importances = pd.DataFrame(bogo_2.feature_importances_,\n",
        "                                   index = features.columns,\n",
        "                                    columns=['importance']).sort_values('importance',ascending=False)\n",
        "feature_importances.plot.bar()\n",
        "plt.title('Best BOGO model feature importance')\n",
        "plt.show()\n",
        "\n",
        "#discount model\n",
        "feature_importances = pd.DataFrame(discount_2.feature_importances_,\n",
        "                                   index = features.columns,\n",
        "                                    columns=['importance']).sort_values('importance',ascending=False)\n",
        "feature_importances.plot.bar()\n",
        "plt.title('Best discount model feature importance')\n",
        "plt.show()\n",
        "\n",
        "#informational model\n",
        "drop_cols_prep=['person','offer_id','offer_responded','offer_type']\n",
        "features,target=data_prep(informational_offer,drop_cols_prep)\n",
        "#print feature importance\n",
        "feature_importances = pd.DataFrame(info_2.feature_importances_,\n",
        "                                   index = features.columns,\n",
        "                                    columns=['importance']).sort_values('importance',ascending=False)\n",
        "feature_importances.plot.bar()\n",
        "plt.title('Best info model feature importance')\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAFbCAYAAAAurs6zAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzt3Xe4XFW5x/HvLwkQOgIRgRACGIIoxZhEBKRIEZRiQYqAAiKXS7Nc0WAXe0UvIk0URZCiAhFRECV0JEFCCTVggIBXASmhJ/DeP9aaZDKZc86c2Xty5uz8Ps9znnP2zJ53rzPlnbXXXkURgZmZVcuQgS6AmZmVz8ndzKyCnNzNzCrIyd3MrIKc3M3MKsjJ3cysgpzcrRIkzZK0Ywv7jZYUkob1cP9YSbdImiPpmPJL2jmSRkl6VtLQgS6LDTwn95LlJPNC/pA9KekPktYpKW6PyUvSdpJezcd9VtIjkr7SsI8kHSvpvlzGhyR9S9IyDfuNl3RJLv9Tku6U9HVJr6nbZ6SksyU9Iek5STdJ2q3o/9kFPg1MiYgVI+J/iwSSNEXSoSWVq08R8VBErBARryyuY/akry9R6zwn987YPSJWANYE/gWcuJiO+2j+cK8AbA18RNJ76u7/X+Aw4EPAisCuwDuA82s7SNoSmAJcB2wUEasAuwDzgM3yPqsC1wIvA28EVgdOAM6RtFcn/8HFYF1gxkAXAmCwJsbBWu7KiQj/lPgDzAJ2rNt+F3Bv3fYywPeAh0iJ/xRg2Xzf6sAlwFPAf4BrSF/AZwGvAi8AzwKfbnLc7YDZDbedD3w2/z0GeAWY2LDPOsBLwDvy9rXAiX38j18F7gCGNNz+GeBBQE0eMxoI4GDgYeBJ4HBgAnBb/p9/XLf/EODzOd6/gV8CK9fdf2C+7wngc/XPe37sJOD+fP/5wKoN5RjWpIx/zc/Ri/l53rCP1+s1+fV6LP8/lwAj831fb4j142bHJn2RHpr/Poj0pXpCfv2/lm8/BLgrH+MyYN0eXpeF4ufYXwOuz2X4PbAacDbwDDAVGF33+ACOAR4AHge+W3uNe3s96o77kfw8XZ1/Rz7us8DbgA3yc/xEjn82sErDZ+dTpPfD08B5wPC6+/cEpuey3w/skm9fGTgD+CfwSP6fhw50LhjonwEvQNV+WDjJLAf8Avhl3f0/BCYDq5Jqz78Hvpnv+2ZOHkvln7eTEyUNXxpNjrsddcmdlMwfYUHSPhx4sIfHXpWPvTwpIW3Xx/94I/CVJrevlz/QY5vcV0sApwDDgZ1Jie8i4LXA2jlpbJv3PwSYCawPrAD8Djgr37dxThjbkJLvD0hnFrXn/eO5jCPz/acCv24oxyLJPd8/hZxsW3i9VgPen1/nFYELgIt6ibXIsVk0uc8DjgaGAcsC78nPwxvybZ8Hru+h7AvFz7FnkpLqysCdwL3AjjnWL4Gf1z0+gCvz/zoq73toC69H7bi/JL2Hlu3hf309sFN+TUaQvgR+2PDZuQlYK5fhLuDwfN9EUsLfifRFszbpzBLSe+jUfOzX5hj/NdC5YKB/BrwAVfvJb9BnSTXRecCjwCb5PgHPARvU7f824B/57+OBi4HX9xC3r+T+aj7uM/mD9Ttg6Xz/54Ebe3jsucDppGQYtQ9Nvu87OeZzwOfzbTNrH7qGOMPz47dqcl/tw7523W1PAPvUbf8W+Hj++y/AEXX3jQXm5qT0ReDcuvuWJzUR1ZL7XcAOdfevWffYRZJOQzmnsCCh9fp6NXns5sCTzWI1PAe9JfeHGmL+EfhI3fYQ4Hma1N4b4+fYn6u7//vAH+u2dwem120HuTact48A/tLC61E77vq9/a9Nyvse4JaG9/gBDe+9U/LfpwInNImxBunMc9m62/YDruzE53sw/bjNvTPeE6mtehngKOAqSa8j1VaWA27OFyqfAv6Ub4d0GjwTuFzSA5Im9fO4j0bEKhGxErAKqRnnF/m+x0lJrpk18/1Pkr4g5u8XEZ/O/8uFpA9yb7HWrLu/J/+q+/uFJtsr5L/XIjUB1DyYj79Gvu/hujI+R/qiqFkXuLDuOb6LdEayRi/laqbX10vScpJOlfSgpGdINdFVCvZWebhhe13gR3XH/w/pS2ftFuO1+nw3O/6DpOcaen89eir7QiS9VtK5+WL/M8CvSE2R9f6v7u/n68q3DqkpptG6pLPcf9Y9R6eSavBLNCf3DoqIVyLid6TEsjUp6b0AvDEn4VUiYuVIF0CJiDkR8T8RsT6pVvVJSTvUwvXz2E8D5+Q4kNo615E0sX6/3JNnC1IN7Tngb8D7+gh/BfB+SY3vn71JH/B7+1PWHjxK+uDWjCKdCf2L1LY6vweSpOVITSQ1DwO71j3Hq0TE8Ih4pJ9l6PX1Av6HVIN9a/5C3aZWpPy78TV7Lv9eru621zXs0/iYh0lNDPX/y7IRcX0//5dW1ffsGkV6HaD316Mmevi75pv59k3z83UAC56rvjxMal5qdvtLwOp1z89KEfHGFuNWlpN7B+Wuh3uSLrzdFRGvkpo/TpD02rzP2pLemf/eTdLrJYnUtPJK/oH0IVq/H8deAdiX3PMjIu4ltXefLWkLSUMlvZHUFHJFRFyRH/pp4BBJk+rKOJLUnl5zArAScIak10kaLmk/0oXNYyOfGxf0a+ATktbL/8s3gPMiYh7wG2A3SVtLWprUnFX/Xj4F+LqkdXP5R+TXoV/6er1I7ewvAE/lHkRfagix0GsWEY+RroMckJ//Q2iesOqdAhyXXyskrSzpA/39X/rhWEmvyV/6HyNd1ITeX49mHiOdBda/Z1ckN1lKWhs4th/lOgM4WNIOkobk12GjiPgncDnwfUkr5fs2kLRtP2JXkpN7Z/xe0rOkBP114MMRUete9xlS08uN+dT0ClLtD9JF0CtIH4AbgJ9ExJR83zeBz+dTz0/1cNy1av3cSafNqwL7191/FPBT0unws6Qmhimki4IARMS1pO6R2wD31jVFTCF36YyIJ0hnIsNJF+meAD4JHBgRtWRQ1M9IvYSuBv5Buvh6dD7+DOBI0pnJP0nNSbPrHvsj0kXQyyXNIV1cfWub5ejt9foh6eLh4/kYf2p47I+AvfJ4gVqf+Y+SktoTpG6kvdbAI+JC4NvAufn4d5C6sHbKxcDNpF4pfyAlVejl9WgmIp4nvfevy+/ZLYCvAONIF0b/QLom1JKIuInU0+qE/PirWHAm8SFgadJ78UnSl39PTZBLDJVTyTKzwU5SAGMiYuZAl8WKc83dzKyCnNzNzCrIzTJmZhXkmruZWQU5uZuZVdCAzd62+uqrx+jRowfq8GZmg9LNN9/8eESM6Gu/AUvuo0ePZtq0aQN1eDOzQUnSg33v1WKzjKRdJN0jaWZP851I2jsv6jBD0jn9KayZmZWrz5p7ngTpJNJUm7OBqZImR8SddfuMAY4jzQb4ZG2otpmZDYxWau4TgZkR8UBEvEyaHrZxno6PAidFxJMAEfHvcotpZmb90Uqb+9osPJXnbBadp2NDAEnXAUOBL0dE4zwbZjaIzZ07l9mzZ/Piiy8OdFGWCMOHD2fkyJEstdRSbT2+leTebErOxpFPw0iTXm1HWvDhGklvioinFgokHUZaw5NRo0b1u7BmNnBmz57NiiuuyOjRo0kTl1qnRARPPPEEs2fPZr311uv7AU200iwzm4XneB7Jgjme6/e5OCLmRsQ/gHtIyb6xwKdFxPiIGD9iRJ89ecysi7z44ousttpqTuyLgSRWW221QmdJrST3qcCYPI/z0qQ5wic37HMRsH0u1OqkZpoH2i6VmXUlJ/bFp+hz3Wdyz5PxH0Vadf0u4PyImCHpeEl75N0uA56QdCdpgd1j85zfZmal2XLLLRfr8WbNmsU55wzOnt0tDWKKiEuBSxtu+2Ld30FarOGTRQs0etIfWt531rfeXfRwZtam/nxWW9HK5/n66zu1uuCi5s2bNz+5f/CDH1xsxy2L55Yxs0FjhRXS8rVTpkxh2223Ze+992bDDTdk0qRJnH322UycOJFNNtmE++9Pa2kfdNBBHH744bz97W9nww035JJLLgHS9YODDz6YTTbZhDe/+c1ceeWVAJx55pl84AMfYPfdd2fnnXdm0qRJXHPNNWy++eaccMIJzJo1i7e//e2MGzeOcePGzf+ymTJlCttttx177bUXG220Efvvvz+1GXenTp3KlltuyWabbcbEiROZM2cOr7zyCsceeywTJkxg00035dRTTy39uRqw6QfMzIq49dZbueuuu1h11VVZf/31OfTQQ7npppv40Y9+xIknnsgPf/hDIDWtXHXVVdx///1sv/32zJw5k5NOOgmA22+/nbvvvpudd96Ze+9N67rfcMMN3Hbbbay66qpMmTKF733ve/O/FJ5//nn+/Oc/M3z4cO677z7222+/+dOo3HLLLcyYMYO11lqLrbbaiuuuu46JEyeyzz77cN555zFhwgSeeeYZll12Wc444wxWXnllpk6dyksvvcRWW23Fzjvv3HbPmGac3M1sUJowYQJrrpmWSt1ggw3YeeedAdhkk03m18QB9t57b4YMGcKYMWNYf/31ufvuu7n22ms5+ui0BOxGG23EuuuuOz+577TTTqy66qpNjzl37lyOOuoopk+fztChQ+c/BmDixImMHDkSgM0335xZs2ax8sors+aaazJhwgQAVlppJQAuv/xybrvtNn7zm98A8PTTT3Pfffc5uZuZLbPMMvP/HjJkyPztIUOGMG/evPn3NfY6kURvixQtv/zyPd53wgknsMYaa3Drrbfy6quvMnz48KblGTp0KPPmzSMimvZ6iQhOPPFE3vnOd/byHxbjNnczq7QLLriAV199lfvvv58HHniAsWPHss0223D22WcDcO+99/LQQw8xduzYRR674oorMmfOnPnbTz/9NGuuuSZDhgzhrLPO4pVXXun12BtttBGPPvooU6dOBWDOnDnMmzePd77znZx88snMnTt3fhmee+65sv5lwDV3M6u4sWPHsu222/Kvf/2LU045heHDh3PEEUdw+OGHs8kmmzBs2DDOPPPMhWreNZtuuinDhg1js80246CDDuKII47g/e9/PxdccAHbb799r7V8gKWXXprzzjuPo48+mhdeeIFll12WK664gkMPPZRZs2Yxbtw4IoIRI0Zw0UUXlfp/D9gaquPHj49m87m7K6RZd7rrrrt4wxveMNDF6JeDDjqI3Xbbjb322mugi9KWZs+5pJsjYnxfj3WzjJlZBblZxswq68wzzxzoIgwY19zNzCrIyd3MWjZQ1+iWREWfayd3M2vJ8OHDeeKJJ5zgF4PafO71/ej7y23uZtaSkSNHMnv2bB577LGBLsoSobYSU7uc3M2sJUsttVSpw+Ots9wsY2ZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFdRScpe0i6R7JM2UNKnJ/QdJekzS9PxzaPlFNTOzVvU5K6SkocBJwE7AbGCqpMkRcWfDrudFxFEdKKOZmfVTKzX3icDMiHggIl4GzgX27GyxzMysiFaS+9rAw3Xbs/Ntjd4v6TZJv5G0TimlMzOztrSS3NXktsZ1tn4PjI6ITYErgF80DSQdJmmapGlezcXMrHNaSe6zgfqa+Ejg0fodIuKJiHgpb54OvKVZoIg4LSLGR8T4ESNGtFNeMzNrQSvJfSowRtJ6kpYG9gUm1+8gac26zT2Au8oropmZ9VefvWUiYp6ko4DLgKHAzyJihqTjgWkRMRk4RtIewDzgP8BBHSyzmZn1oaUFsiPiUuDShtu+WPf3ccBx5RbNzMza5RGqZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQU7uZmYV5ORuZlZBLSV3SbtIukfSTEmTetlvL0khaXx5RTQzs/7qM7lLGgqcBOwKbAzsJ2njJvutCBwD/K3sQpqZWf+0UnOfCMyMiAci4mXgXGDPJvt9FfgO8GKJ5TMzsza0ktzXBh6u256db5tP0puBdSLikhLLZmZmbWoluavJbTH/TmkIcALwP30Gkg6TNE3StMcee6z1UpqZWb+0ktxnA+vUbY8EHq3bXhF4EzBF0ixgC2Bys4uqEXFaRIyPiPEjRoxov9RmZtarVpL7VGCMpPUkLQ3sC0yu3RkRT0fE6hExOiJGAzcCe0TEtI6U2MzM+tRnco+IecBRwGXAXcD5ETFD0vGS9uh0Ac3MrP+GtbJTRFwKXNpw2xd72He74sUyM7MiPELVzKyCnNzNzCrIyd3MrIKc3M3MKsjJ3cysgpzczcwqyMndzKyCnNzNzCrIyd3MrIKc3M3MKsjJ3cysgpzczcwqyMndzKyCnNzNzCrIyd3MrIKc3M3MKsjJ3cysgpzczcwqyMndzKyCnNzNzCrIyd3MrIKc3M3MKsjJ3cysgpzczcwqyMndzKyCnNzNzCrIyd3MrIJaSu6SdpF0j6SZkiY1uf9wSbdLmi7pWkkbl19UMzNrVZ/JXdJQ4CRgV2BjYL8myfuciNgkIjYHvgP8oPSSmplZy1qpuU8EZkbEAxHxMnAusGf9DhHxTN3m8kCUV0QzM+uvYS3sszbwcN32bOCtjTtJOhL4JLA08I5mgSQdBhwGMGrUqP6W1czMWtRKzV1NblukZh4RJ0XEBsBngM83CxQRp0XE+IgYP2LEiP6V1MzMWtZKcp8NrFO3PRJ4tJf9zwXeU6RQZmZWTCvJfSowRtJ6kpYG9gUm1+8gaUzd5ruB+8oropmZ9Vefbe4RMU/SUcBlwFDgZxExQ9LxwLSImAwcJWlHYC7wJPDhThbazMx618oFVSLiUuDShtu+WPf3x0oul5mZFeARqmZmFeTkbmZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQS0ld0m7SLpH0kxJk5rc/0lJd0q6TdJfJK1bflHNzKxVfSZ3SUOBk4BdgY2B/SRt3LDbLcD4iNgU+A3wnbILamZmrWul5j4RmBkRD0TEy8C5wJ71O0TElRHxfN68ERhZbjHNzKw/WknuawMP123Pzrf15CPAH4sUyszMihnWwj5qcls03VE6ABgPbNvD/YcBhwGMGjWqxSKamVl/tVJznw2sU7c9Eni0cSdJOwKfA/aIiJeaBYqI0yJifESMHzFiRDvlNTOzFrSS3KcCYyStJ2lpYF9gcv0Okt4MnEpK7P8uv5hmZtYffSb3iJgHHAVcBtwFnB8RMyQdL2mPvNt3gRWACyRNlzS5h3BmZrYYtNLmTkRcClzacNsX6/7eseRymZlZAR6hamZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkFO7mZmFeTkbmZWQU7uZmYV5ORuZlZBTu5mZhXk5G5mVkHDBroAi8voSX9oed9Z33p3B0tiZtZ5rrmbmVWQk7uZWQW1lNwl7SLpHkkzJU1qcv82kv4uaZ6kvcovppmZ9UefyV3SUOAkYFdgY2A/SRs37PYQcBBwTtkFNDOz/mvlgupEYGZEPAAg6VxgT+DO2g4RMSvf92oHymhmZv3USrPM2sDDdduz821mZtalWknuanJbtHMwSYdJmiZp2mOPPdZOCDMza0ErzTKzgXXqtkcCj7ZzsIg4DTgNYPz48W19QXQb9583s27USnKfCoyRtB7wCLAv8MGOlsoAf3GYWfv6bJaJiHnAUcBlwF3A+RExQ9LxkvYAkDRB0mzgA8CpkmZ0stBmZta7lqYfiIhLgUsbbvti3d9TSc01ZmbWBTxC1cysgpzczcwqyMndzKyCnNzNzCrIyd3MrIKc3M3MKsjJ3cysgpzczcwqyMndzKyCnNzNzCrIyd3MrIKc3M3MKsjJ3cysgpzczcwqyMndzKyCnNzNzCrIyd3MrIKc3M3MKsjJ3cysgpzczcwqqKUFsq1aRk/6Q8v7zvrWuztYEjPrFNfczcwqyMndzKyCnNzNzCrIyd3MrIJ8QdVK08kLtb4IbNY/Tu62RPOXhlVVS8ld0i7Aj4ChwE8j4lsN9y8D/BJ4C/AEsE9EzCq3qGaDh780bKD1mdwlDQVOAnYCZgNTJU2OiDvrdvsI8GREvF7SvsC3gX06UWCzJV2nvjj8hVQtrdTcJwIzI+IBAEnnAnsC9cl9T+DL+e/fAD+WpIiIEstqZoNQN3wZ9Td2Faiv/CtpL2CXiDg0bx8IvDUijqrb5468z+y8fX/e5/GGWIcBh+XNscA9LZZzdeDxPvdqT6diD7a4nYztuJ2PPdjidjL2YIvb39jrRsSIvnZqpeauJrc1fiO0sg8RcRpwWgvHXDi4NC0ixvf3cQMZe7DF7WRsx+187MEWt5OxB1vcTsVupZ/7bGCduu2RwKM97SNpGLAy8J8yCmhmZv3XSnKfCoyRtJ6kpYF9gckN+0wGPpz/3gv4q9vbzcwGTp/NMhExT9JRwGWkrpA/i4gZko4HpkXEZOAM4CxJM0k19n1LLme/m3K6IPZgi9vJ2I7b+diDLW4nYw+2uB2J3ecFVTMzG3w8t4yZWQU5uZuZVZCTu5lZBS2RyV3Seq3cViD+8mXFMrPWSFpW0tiBLke36OoLqpI+APwpIuZI+jwwDvhaRPy9YNy/R8S4httujoi3FIy7JfBTYIWIGCVpM+C/IuKIgnE3BE4G1oiIN0naFNgjIr5WIObtNBloVhMRm7YZd1xv97f72kn6RkR8Nv+9U0T8uZ04vcQ/Cjg7Ip4sKd7v6f353aNA7FV7uz8iCo8xkbQ2sC51Peoi4uqCMdcnTUD4NuBV4AbgE7WpTQrG3h34HrB0RKwnaXPg+Haf506+fjn+a4EjgTfm49wJ/CQi/lUkbr1un/L3CxFxgaStgXeSXryTgbe2E0zSRqQnc2VJ76u7ayVgeNHCAieQyjkZICJulbRNCXFPB44FTs1xb5N0DtB2cgd2y7+PzL/Pyr/3B54vEHcaMAN4LG/Xj14O4B1txt0F+Gz++9tAqckdeB1pUry/Az8DLis4VuN75RSrqZtJz6WAUcCT+e9VgIeAQmehkmoT/90JvJJvDqBQcgfOIU1C+N68vS/wa9r8PDf4MmkerCkAETFd0ugC8Tr2+knaivRcnEmaTVekiuvfJO0fEdeVcZxuT+61N9a7gZMj4mJJXy4Qbywpqa0C7F53+xzgowXizhcRD0sLzcbwSk/79sNyEXFTQ9x5RQJGxIOQ3mgRsVXdXZMkXQcc32bo/wHeD7wAnAtcGBHPFinr4hARn5f0BWBn4GDS5HfnA2dExP1txLuq7DLWxV4PQNIpwOSIuDRv7wrsWMIh3gOMjYiXSohVTxFxVt32r/IZUxnmRcTTDZ+RtnXy9QO+D7wnIm6pu+1iSReSKnBlfNl1fXJ/RNKppDfst/O88W1fJ4iIi0lP4tsi4oayClnn4dw0E3k07zHAXSXEfVzSBuTTxDyZ2z9LiAuwvKStI+LaHHtLoO1rBhFxAnBCvoaxH/AXSQ8C34iI6QXK+VpJnyTVcmp/1x/3BwVi12KEpP8D/o/05fka4DeS/hwRn24npqQxwDeBjak7O4yI9YuWF5gQEYfXxfyjpK+WEPcBYCmglORe14x0paRJpC/9IJ0d9G9qx57dIemDwND8nB8DXN9uMEnnR8TePTVftttsma3UkNhrMadLWrFA3IV0e5v7cqTT8dsj4j5JawKbRMTlBeOOINXUR7Nwm+IhBeOuTmpT3JGUhC4HPhYRTxSMuz5pBNuWpFPwfwAHlLEgiqS3kJohVs43PQUcUvS6Ro79RtKp94HApyPi/AKxvtTb/RHxlXZj5/jHkKbQeJx03eSiiJgraQhwX0Rs0Gbca4EvkZrsdiedFSgiev1/Wox9GXAN8CtSAjoA2CYi3lkw7m+BzYC/UJfgI+KYNuP9gwXNSI2ijC+6nCs+RzrzEmlE/Vcj4sU2460ZEf+UtG6z+2tnvm3GvgvYsvH6Tv4SvD4iNmo39kLxujy5fw/4eUTMKDnu9aQPxc3UNZtExG/LPE7Zci+cIRExpwOxVyK9H54uGGd9UkLfE3iYVEu7pN0PWRvHPy4ivtnG444nNcEs8qGV9IaIaOsMrHahXtLtEbFJvu2aiHh7O/EaYq9K+uLYhgVt4scXvaAq6cPNbo+IXxSJa0me+vyjwKeAWiXqLaRrST+LiFNLOU6XJ/dDSTWdYcDPgV8XTT457vSI2LxonCZx/7fJzU+T5uC5uEDcVYAPseiZRls1qRzzgIj4VWPzRl3stpo5JL0K3AZcDDxDwyltGc0nfRx/kZ5QLT7urIg4sK/b2oh7HfB20iI2fwUeAb4VEYW67CmtkPatiDi2SJxe4i8NbJg374mIuSXEXA74JDAqIg7LzSdjI+KSAjE73atlC+BE4A3A0qT5tZ6LiJUKxt0N+DQL95b5bkT8vkjcel3d5h4RPwV+mvuuHgzclj8sp0fElQVCXyLpXbULUSUaDmwEXJC330/qOfIRSdtHxMfbjHspcCNwO6kLWRlq7eqltfFlx7Pgw7ZCybFb0e4VtTcuFCQlz0JdY7OPA8uR2oC/Suot1LRm3B8R8UpuUiudpO2AXwCzSM/nOpI+XLQrJKmCdjOpeRHSVOEXAG0ndzrbKwngx6Qz0QuA8aRK1uuLBs1faL3+3+2ehc5/fDfX3GH+h2w3UnJfBzgf2Jr07dnW7JOS5pCS20vAXNIbOEr4Nv4rsHNEzMvbw0jt7juRrhts3Gbctmqj3azoG7eXuP16riQdR+piuSwLuoAjf3SBAAAYIklEQVQKeBk4LSKOK7uMZZH0fWAMKfE8V7s9In5XMO7NwAcj4p68vSHprLnoOJBpETFe0i0R8eZ8260RsVmRuJ1UV+bbahdRJV0fEVv29dgSjl3oc9/VNXdJPwD2IF3Y+UZE3JTv+rakVpfoW0RElF1brVmb9KVRazpaHlgr17KK9Dw4S9JHSd/09Re42m5b7aEJab4iTT4t+gCpF0nZ+lVzz18w35T0zTITuaQfRsTHe2o2KNpckK0KPMHCYwcCKJTcgaVqiR0gIu6VtFTBmAAvS1qWBb2+NqBgj5xeerXUKmxFerUAPJ+bqKZL+g6pl9riGoFeqF9nVyd34A7g8xHRbFDNxHaDSnovaUGRp/P2KsB2EXFRuzGz75DeBFNIL8w2wDfyhdArCsR9GfguqTdA7Q0cQJFeBjcXeGwZ+v3GzWdxx+Tulj25oJf7msWs1YwuUJPRtQV6DdX6c3es2SAiDu5Q6GmSzmDhgW1lvF++BPyJ1MxzNrAVcFDBmB/Lv3frda/2HUhqZz8K+ASp9eD9HTpWo0LNKoOhWeY1pFPP+j7CRYdBL3JBtf5UsWDstUhviLtJ3/CzSyhv0wXHB7MCFz6nRMR2JZajt2s3ERHtjqitxV8eeCEiXs3bQ4Fleqiw9Df2cOAjpOsF9Z+Pol16lyGNXN6a9CV8NWlofNFa9lmk60YvkPrS/63M97Sk15EqfQFMjYj/Kyv2QCiak7q65p57y3yMtG7rdGAL0nwUhT5wNB8IVfi56GB5Z1BsSoAe5T7/n2HRQTZFy9znodt83HWSfgycx8LtzG3VsCNi+zbL0aq/kMY91EbpLku6DlNGm+1ZpErEO0kXsvenhEFzOYn/IP+U6eekL4ydSGed0yVdHRE/Kho4f/a+SOqRJOBEScdHxM8Kxt2NdCG8Ns9OWdfnSj8LXeQY3Vxzz+1oE4AbI2JzpblhvhIR+xSM+zPSYJ2TSN/yRwOviYiDurS8F5JqZ1dSwqCShtiXkxLlp4DDST05HouIzxSI2ecbV9JnI+IbbcRuVtMuo4b9oWa3R8QvC8ZtdpZYSlfcWs2udrEvt4tf1u5z0Uv7NVB4VGbtGENJn5HtSe+3F6KEQTv5GtyWkQcMSlqNNCCoaJfTmcD7SB0iSk2WZZ+FNurqmjvwYkS8KAlJy0TE3SpnSs+jgS+QkhqkmtTnS4jbqfJelH86YbWIOEPSxyLNp3GVpELzauQLyHuSRmX2tE+/E3t+XKdq2hPq/h4O7EAaYFIouQPPSRpXO7PI3RdfKBizptb3/ClJbyJNmzC6QLyOtl9L+gupqfIG0iDCCRHx75LCzybNEVUzhzSIrqiHgTvKTuxZqWehjbo9uc/OFzsvAv4s6Ung0aJBI+I5YFJP90s6MSKObiN0p8r7i04MKslqcf4p6d2k8o4sIW5H3riS1gC+QeqFtKukjYG3RcQZReI2vt6SVmbBBcUiPk66WFt7H6xJmlOlDKfla1JfIM1EukL+uy0RUZuv6IjGMzelmSLbPpvLbiONHXgTqUfZU5JuiIi2v+y0YBDeI6RZFS8mnXXsCdzU4wNb92ng0lzhqT9rLqPJqtY0Vz9JX5GZUxfS1c0y9SRtS5r/5E8R8XKHj1W4X3mZ5VWTQSVAGYNKam2K1+SYJ5KmP/5KREwuGLdTzSd/JLXdfi4iNstjCW6JPLS/LLmJ47aIeENJscaSXru7S/xi7ohm7//6ft4lxF+BNG7lU8DrImKZArE6PefQ5aTrJQsNICwad3HoyuSuxbAYQR/H76pBQ+rQoJIca0REPNb3nt1B0tSImKCFB8IUbsNu6I8+hHSB+fyI6PEMr8W4SwH/TeoWC2m+8VPLSPC5F9WNpC/nqyPizoLx/hs4gnSxs36a4xWB6yLigILxjyJNxfAW4EFSL5xrIuKvReI2HGNFUiWilGmmlQcxlRGrSeyOnIXWdGuzTEcXIxiEOjWoBOB6pVn7zgN+F+WtRNSpN+5z+WJZbSDMFiwYNFZEfX/0ecCDETG7hLgnk6bP/UnePjDfdmgJsTcmzf39duB7+QL+rRHx3t4f1qNzgD+SBpfVf6nNKalCtSypB87NkUdxlyVfcziLNLALSY8DH4rikw5eIWnnKDgTbQ/OJJ+F5u17SZ/DUpI7EdG1P8ApwLvqtncFvr8YjnvLQP/vDeX5WX7Bt8s/p5Nmyywr/kTSh+4B0ijYA0qI+Udgb1KygVSRuL2EuOOA60gJ/br8gdi0hLjrAcPrtpcFRpcQ99ZWbmsz9jDSknWT8ut2A+msoKz3xWtJlatRpMm+SonbiR/S3O3b121vR+otUzTuHNLMsS+QJsKbAzxTUpmn5t+31N02vbTnZKBflD7++Zub3DZtMRz3oIH+3xvKswxpNr3fAReSRsot04HjrE7qHfJKCbE69sbNSe2NpAtzS5UUcxpp/c3a9tK1/6Fg3L8DG9Rtrw/8vaQyPw/8jXSBdrUS3we7A/eRLoT/g9TWPKPs91uZP536EiU10R0IfDFvjyINKCyjzFOA1WrvB9K4mKvKek66tVmm5nGlhbHrFyNoe+ELtTg9aESc2e4xOmQY8KPIV+hroxzLCKw0j/t7STPfbUD68mh7aoc6pTafaOE1b+ttKIkoOFkWMCzqLnxHxMu5h1JRnyKtQFRbBHo06WJiGfYjDQo6AjhUaZ2CqyPiLwXjfo2UaK6I1I9++3ysbvaA0jKJtR5OB5C+mIo6ifTl9g5Sr5Y5wG9ZuOtsuz5J6uW0gdJstyOAvUqIC3Rvm3vNfqT5KC5kwWIERd5ktXbV95EWRP5V3XFmFYjbaZ0c5Xgrqevm8VHu0oNlv3Fra96+lvR/1y7CbU+qARVN7o9J2iNyL6HcT7+MofGrkc4wRpO6521JOdcIiAXLRm5EarL8OKnr3rIFQ8+NiCckDZE0JCKuzF0hu9khwFdIibc2ZcJBJcR9a0SMk3QLQEQ8WdKXPhHx99yrrtaTqswuzt2d3CNdxPlYT/f3tz965EVvJX01Irapu+v3kgp3K+yg4VF39T8inlVa+KAM60c+JyxT2W/cyJNkSboE2Dhyn2ylpRdPKqHIhwNn5775Ig1eaTpqtZ++EBEX5DOknUiLI59MCYsgKy2Htzkwk9Rj5kOkZpqinsrdFa8mPSf/puCC7IvBBqTuvENIeW0HUm27aPfNuflMuXYGOoKCayoshrNQoMuTewu2avNxIyStHxEPACgt5jyivGKVrpOjHMdI+hSLrvLU7hD2Tr9xR8eCwTYA/2LB4K62RcT9wBY5qSnKW8qwtozju4FTIuJiSV8uKfa3SO21r/S5Z//sSXp/fYI0X83KLDzQphudTWoCu4PyFrQB+F9Sy8FrJX2ddPZZdDR7p89CgcGf3Nv1CWBKQzvofw1ccfrUyVGOF5B6Jf2UuvVkC+j0G3eK0sLQvybVpvYlzbnTFvWw3KCU5jWL4iMRH5F0KqlZ7dtKMy42m7iuHTOA4ySVuWzdUODiiNiRlCQHy7qpj0WJS9TVRMTZeZzJDqQzuvdEm+vp1sXs9FkosIQm94j4U/4g1CYsujsKTmfaSRExNberdmKU47yIOLmkWB1/40bEUfnsoLbA9GkRcWGBkLXmrU4t4LI3sAvwvYh4Kj8PZa17WvqydZHmBXpe0spRwnrFi9GXJP2UdH2qfpqAwrXgiLibNPtm2TpyFloz2JN7W9PGasFCvetGxEcljZFUqMazGExgQdPJm3MTR9FJrSBdbziCdOpZyipPWcfeuPkDW8qpK6mtFuDOiCg0xWozkeZt/13d9j9Jq/mUYYOI2EfSfjn2C6qdchTzInC7pD+z8LxAnV6dq4iDSZW1pVjQLFPGqlSdVOpZaKNBkdzzxaho0g7a7jzQtRrP2/J2GQv1dozSIgcbkOaIrzWdBMVnLIQFizXX1yaLrvIEHXrj5lr7t0nNPqL4/Nrvyt1tj6Pg/NkDoPRl67I/5J/BZLMoeX6hTuvAWehCunJumRpJ40mJeEXSh/gp4JCIKLTklwbZQr2S7iI1cXTvi9VEwxv36jLeuErza+9etN2zLt53gcNIU9HWL4hSyqIMnZJr6AeSVmLamNQ1divSALwpA1i0ASHpdOCEKDi/TpV0e3K/DTgyIq7J21uTlvsq1L0pD/bYgTQZ0rhc4/l1RJQxeKd0ki4gLX5R1uk8kt4REX/tqXdLWd2xyibpuohot5dUs3jLRMRLki6OiD3Lirs45At9O5MGHIm0SEzhvvl5rqFmi3UUPZvrmFwB2oA0cOklFnw5lzKTZSd04Cx0Id3eLDOnltgBIuJaSWV0Ufsy5S/U20mrA3dKuomF28X3KBBzW1JPlt2b3Fe4rbKDb9xpks4jDbwq48LZDaT5ap4pWK6BcCNpnELZTSj1syAOBz5AnpCri+0y0AVow3co8Sy0UbfX3E8g9WaotdvuQ5oh8rdQbOGHPDS+1BpPp+TBQIuoDcrq8LE/HBH97g5XdvNJXdyfN7k5os1FoSXdAXyXtP7mIr1YuvUMBkDSnaSL1A+SLnx2rLYq6dqI2LrsuEuyss9CF4nf5cm9IyvTS5pM+sKYHGlVJuuB2pzbvtNv3LLkpr79SV0WGxcoaftLY3GQtG6z2yPiwYJx61/vIaSa/H936zWpwUrSj0jToJR1Frpw/G5O7p2Sa8L7kEYN3kSaQ/mSiHhxQAvWoFZbyk1R9S/UYrvYV3/RuZ+P68gbV2mhkpOBNSLiTZI2BfaIiK8VjPuRKGmRhMEuV6pq77d5pHmXvhcR9w5YoSqo7LPQReJ3Y3LvadRgTQmjBmvHGUqaf+KjwC7d2jNiIBWouXfkjau0luWxpHnLaz2d7oiIN7UZb1BeWO6Eus+bWLBYDvnv0j53tnh06wXV5fPvTo0aJPcP3p1Ugx/H4Blmvbi1NSimNlK1A5aLiJsaxuoUmdRqGxZcWK4ltPrfS0xyZ8HnbSxp0NzFpOdhd9IkYlaiTp2F1nRlco+IU/PvjixCm3tbvJXUY+YkYEpElDnZ0KAhab2I+Ecvt13XZtxOvXEfz11XawN39qLYiM85ucZ6B01qq0uS2udNaVHocbVBg3mis8E2wGswOJ18FgoQEbdJOoc0n35hXZnca5Sm1/woi85YWLRN6uekBafLnk1vMPot6cyl3m9IixgTEUe1GbdTb9wjgdOAjSQ9QurXvH+BeCvk366tLjAKeLlu+2XSZ9DKVfZZ6EK6OrmTPmjXAFdQwoyFtfZVUvfKPRun4VjC2lc3Ii1Vt3JDe/NKpL7NRXXqjfse4FLSVAZDSF0Ad5R0c0RM728w11abOgu4SVJtkZz34mbLTij7LHQh3Z7cl4uIz5QYr6MDdwaZscBuwCos/HzMIZ0tFdWpN+74/DOZVMPeH5gKHC7pgoj4TptxXVvNIuLrkv7IgqkjDo6IWwayTBVV9lnoQrqyt0yNpK+RVjC/dKDLUlWS3hblLq9Xi7s+6Y27JWng2T+A/Uvog30Z8P7IK1MpLa7xG1Lt8uaI2LjNuJ8j9XWvr62eFxHfLFJes57U9U5algVnoU+T3sf9PgtdJH43Jve6ft0i9Zx5CZhLSf27Ja0BfANYKyJ2lbQx8LYlsZ9zp65rdOqNm+cQ2SzyYtZKi19Mj4g3tNsnvy72OBae6My1VeuYfA2q/iz03aSz0I2AImehQJc2y0REx7pAZmeSLqp+Lm/fSxrItMQld0q+rlGnU80n5wA3Sro4b+8O/FrS8kChGQHzdBZtT2lh1k+rka7z1M5Cv0Q6C92GNCV5oeTelTX3GklbkWplz0k6gNSr44cR8VDBuFMjYkLDlL/TI2LzEoo9qHTq/+5U80mO9RZga9KXxrURMa2EIpstVp08C4UurbnXORnYTNJmwKdJNeuzSBdGi3guTxxWu9i3BanJYEl0iaR3deC6RuMFyrmkla9ekFRoQYlI8/kXmtPfrAt07CwUur/m/vdI861/EXgkIs5odzh8Q9xxwInAm0iDV0YAe0XEbcVLPbjk6xuduK7xBVItvf6NOxn4PmnFmdJ6BZgNVp08C+325H4VaRTpwaR2qMdIpy2Fl9OSNIwFC07fE+UtOG2Zm0/MBk63J/fXAR8EpkbENZJGAdtFwYWhJR0JnB0RT+Xt1wD7RcRPChd6kJG0TbPbI2JJHZ1pVgldm9zzjI2XRcSOHYi9yEXEMi5gDEaSfl+3ORyYSLrg2dZc+WbWHbr2gmpEvCLpeUkrR0TZFzuHSFLkb7b8RbJ0yccYFCJiodG6ktahYBcsMxt4XZvcsxeB2yX9mTQIBoCIOKZg3MuA8yWdQuoxczipbd9gNulCs5kNYl3bLANp/c5mt0cba3o2xB0C/BewA+li3+XAT5fEWSIlnciC6W2HAJsDsyLigIErlZkV1dXJHeYvqjEqIu4ZDHEHm4Yv0HmkxN7WHO5m1j2GDHQBeiNpd2A6uclE0uZKi1sXjbtHJ+IORvks6NekQUG3ktaUNbNBrquTO/BlUu+NpwDyhFPrlRD3S03iji4h7qAjaTvgPtKKVD8B7u2pe6SZDR7dfkF1XkQ83bDgQxntSM3iLqm+D+xca57Ky+P9mrwSk5kNTt2e3O+Q9EFgqKQxwDHA9V0cdzBaqv66Q0TcK2mpgSyQmRXX7c0yR5OWgnuJVJt8Bvh4yXHPIU0aVkbcwWiapDMkbZd/TseTcpkNel3fWwZA0kqkyazmlBBrKPCtiDi2eMkGvzzN6JEsmAPmauAnEVFo5kYzG1hdndwlTQB+BtQW73gaOCRP+Vok7l89vN7Mqqzbk/ttwJERcU3e3ppUq9y0YNzvA2NIq9vXj3xdkhbIBkDSbsBXgXVJ12BKmfLXzAZWt19QnVNL7AARcW2ef7yoVYEngPraewBLXHIHfgi8D7g9uvmb3sz6pStr7nkxDYADgeVIF1MD2Ad4MiI+19NjSzr+cUvKqveSrgR2iIhXB7osZlaebk3uV/Zyd3S6vbyM1Z4Gi3xd46vAVaTeQwBExA8GrFBmVlhXNstExPYDXIQlaXTT14FnSXO5L5HTHptVUVcm9xpJqwAfIk0NML+sJUz525fuO53pnFUjYueBLoSZlaurkztwKXAjcDuwONuEl6Sa+xWSdo6Iywe6IGZWnq5sc6/pRNt3HsR0TESc0Ms+n42Ib5R53G6Vex8tT2pvn4u7QppVQrcn90+Q2oMvYeGLff8pGHdKRGxXrHTVIWlVUr//4bXbIuKqgSuRmRXV7cn9SNIFv6dY0A4eEbF+wbhfB1YGzmPhQUx/LxJ3MJJ0KPAxYCRpjvstgOsjYocBLZiZFdLtyf1+4K0R8XjJcZt1tex4F8tuJOl2YAJwY0RsLmkj4CsRsc8AF83MCuj2C6ozgOfLDtoFXS27yYsR8aIkJC0TEXdLGjvQhTKzYro9ub8CTM817fo290JdISWtAXwDWCsidpW0MfC2iDijUGkHp9m5y+lFwJ8lPQk8OsBlMrOCur1Z5sPNbs/rfhaJ+0fg58DnImIzScOAWyJikyJxBztJ25KuRfwpIl4e6PKYWfu6OrkDSFoWGFW/WlAJMadGxARJt0TEm/Nt0yNi87KOYWY2kLp6JSZJu5N6cPwpb28uaXIJoZ+TtBq5B46kLUhzxZuZVUK3t7l/GZgITAGIiOmS1ish7ieBycAGkq4DRgB7lRDXzKwrdHtynxcRT0sLzQbQdjuSpA9ExAXAk8C2wFjSiMx7ImJuoZKamXWRrm6WAe6Q9EFgqKQxkk4Eri8Q77j8+7cRMS8iZkTEHU7sZlY1XX1BVdJywOeA2qyFlwFfbXfxZklXAEOBzYFrGu+PiD3aLKqZWVfp9uQ+npTcR7OgCSnaXUNV0tLAOOAs4NDG+z2fiplVRbe3uZ8NfAq4g3Km/D0jIg6UdLoTuZlVWbcn98ci4vclxnuLpHWB/SWdTsO87UVnmzQz6xbd3iyzA7Af8BcWnn7gd23GOwb4b2B94JH6uyhhtkkzs27R7cn9V8BGpAnEas0yERGHFIx7MnAKsE2+6eqIuLVITDOzbtLtzTKbdWi+l7uBXwG/I9Xaz8rt8Cd24FhmZotdt9fcTwdOiIg7S457G2kWyOfy9vLADe32wjEz6zbdXnPfGviwpH+Q2txrbeNFk7BI0wnXvMKStSi2mVVctyf3XToU9+fA3yRdmLffAyyJc7mbWUV1dbNMJ0kaRzozEOmC6i0DXCQzs9IsscndzKzKun3iMDMza4OTu5lZBTm5m5lVkJO7mVkFObmbmVXQ/wO9QJBeBKVGbgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f2cf7fc91d0>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f2cf7fd64a8>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f2cf800d240>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_ybgZ_SuyBLa",
        "colab_type": "text"
      },
      "source": [
        "As shown above, we can see that for all three types of offer, the most important factor that largely affects if the offer will be responded to eventually is the length of membership. That is, the longer the customer as a member of Starbucks, the more likely (s)he will respond to the offer they receive. Then the second and third important factors which affect the possibility of customer's response are age and income which very make sense. Also, the number of offers they received will also affect the response a lot."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Q3n_reQwyBLa",
        "colab_type": "text"
      },
      "source": [
        "## Conclusion & Next steps"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "09YXvw3VyBLa",
        "colab_type": "text"
      },
      "source": [
        "### Conclusion"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mc1tga-YyBLb",
        "colab_type": "text"
      },
      "source": [
        "From the result of the project, it's likely to use machine learning model to predict whether the customer will respond to the offer or not, and the model also shows the main factors such as the length of membership, age, income which highly affect the possibility of customer's responding to the offer.\n",
        "\n",
        "This project is trying to figure out:\n",
        "- What factors mainly affect the usage of the offer from customer? Should the company send out the offer or not?\n",
        "\n",
        "For this question, the feature importance given by three models shows that the length of the membership is the most important factor which affect how customer will make reponse to the offer. \n",
        "\n",
        "For all three models, the top three variables were pretty same - length of membership, income and age. However, income and age switched orders depending on offer type. And for BOGO offer, the count of offer received is the second important factor which decide if the customer will make the transaction.\n",
        "\n",
        "- How possible will a customer open and use the offer sent to them? Is there any common characteristics on the customers who take the offer?\n",
        "\n",
        "In this project, it uses thress separate models to predict if the customer will repond to the offer or not using 'offer_responded' trage variable and finally get the accuracy of 0.83, 0.87, 0.75 for BOGO, discount and informational offer seperatly. Although, the model performance of informational offer is slightly lower but is still acceptable in this phase of research. And for BOGO and discount models, the cost won't be too much to send out the offer even some people doesn't repond to it and it should be fine with this level of model performance."
      ]
    }
  ]
}