From b2797ab40ecfc8f224537e2bde2fe81dffe704f4 Mon Sep 17 00:00:00 2001 From: Peter Min Date: Sat, 23 Feb 2019 10:09:47 -0800 Subject: [PATCH] removed misc cells from notebook --- .DS_Store | Bin 14340 -> 14340 bytes .../Benson_geolocation_final-checkpoint.ipynb | 272 -- .../Benson_geolocation_final.ipynb | 272 -- ...crapcode selenium testing-checkpoint.ipynb | 141 + .../scrap/Scrapcode selenium testing.ipynb | 24 +- 04-Aspect_Based_Opinion_Mining/.DS_Store | Bin 10244 -> 10244 bytes .../01-Build_Model-checkpoint.ipynb | 48 +- .../02-Sentiment_Analysis-checkpoint.ipynb | 107 +- ...int.ipynb => 03 - Plotly-checkpoint.ipynb} | 11 +- .../code/01-Build_Model.ipynb | 64 +- .../code/02-Sentiment_Analysis.ipynb | 107 +- ...ap_code_Plotly.ipynb => 03 - Plotly.ipynb} | 11 +- 05-Uber_demand_forecasting/.DS_Store | Bin 10244 -> 10244 bytes .../02 - GRU_Uber-checkpoint.ipynb | 3036 +++++++++++++++++ .../code/02 - GRU_Uber.ipynb | 2 +- 05-Uber_demand_forecasting/data/.DS_Store | Bin 6148 -> 6148 bytes 16 files changed, 3228 insertions(+), 867 deletions(-) delete mode 100644 01-MTA_Turnstile_EDA/.ipynb_checkpoints/Benson_geolocation_final-checkpoint.ipynb delete mode 100644 01-MTA_Turnstile_EDA/Benson_geolocation_final.ipynb create mode 100644 02-Movie_Opening_Gross_Prediction/scrap/.ipynb_checkpoints/Scrapcode selenium testing-checkpoint.ipynb rename 04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/{03 - Scrap_code_Plotly-checkpoint.ipynb => 03 - Plotly-checkpoint.ipynb} (99%) rename 04-Aspect_Based_Opinion_Mining/code/{03 - Scrap_code_Plotly.ipynb => 03 - Plotly.ipynb} (99%) create mode 100644 05-Uber_demand_forecasting/code/.ipynb_checkpoints/02 - GRU_Uber-checkpoint.ipynb diff --git a/.DS_Store b/.DS_Store index 0655d2113783273e91fc2e3a1ad3d48c99cba7de..d2274a1b4d21353cff4e903eb0ff9ee49f99f77a 100644 GIT binary patch delta 92 zcmZoEXepTBXEU^hRb?8FIflOu(>H^&OdvT_(2o9ieTnix#(7joQuSab;!Tx4^; bgb~YTF~%(B&D;tx92nwh%$wN_{;~rA-dq_v delta 125 zcmZoEXepTBLAU^hRb+{6iPlOu(>H^&OdvQBOn^4xq_bO{rfu|eV^vnUrs5Q8g& qBZCWrFOaTd&|}DDNSQoOP=2!*qZ#vNZiN^Q;uIS(Z)P|6%MJiE@gjx* diff --git a/01-MTA_Turnstile_EDA/.ipynb_checkpoints/Benson_geolocation_final-checkpoint.ipynb b/01-MTA_Turnstile_EDA/.ipynb_checkpoints/Benson_geolocation_final-checkpoint.ipynb deleted file mode 100644 index 12472a2..0000000 --- a/01-MTA_Turnstile_EDA/.ipynb_checkpoints/Benson_geolocation_final-checkpoint.ipynb +++ /dev/null @@ -1,272 +0,0 @@ -{ - "cells": [ - { - "cell_type": "code", - "execution_count": 1, - "metadata": {}, - "outputs": [], - "source": [ - "#Initial Imports\n", - "import pandas as pd\n", - "import numpy as np\n", - "import matplotlib.pyplot as plt\n", - "import seaborn as sns\n", - "import datetime as dt\n", - "import glob\n", - "import os\n", - "import googlemaps" - ] - }, - { - "cell_type": "code", - "execution_count": 2, - "metadata": {}, - "outputs": [], - "source": [ - "# Read NYC turntable for the month of march, from march 3rd to march 30th (4weeks)\n", - "#df = pd.read_csv(\"NYCT180310.csv\")\n", - "path =r'/Users/petermin/metis/01-benson' # use your path\n", - "all_files = glob.glob(os.path.join(path, \"NYCT*.csv\"))\n", - "all_files\n", - "df = pd.concat((pd.read_csv(f) for f in all_files))\n", - "df.columns = df.columns.str.strip()\n", - "df;" - ] - }, - { - "cell_type": "code", - "execution_count": 5, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "array(['59 ST', '5 AV/59 ST', '57 ST-7 AV', '49 ST', 'TIMES SQ-42 ST',\n", - " '34 ST-HERALD SQ', '28 ST', '23 ST', '14 ST-UNION SQ', '8 ST-NYU'],\n", - " dtype=object)" - ] - }, - "execution_count": 5, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "# Get the list of station names in the dataframe\n", - "stations = df[\"STATION\"].unique()\n", - "stations[:10]" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "## Get Station Zipcode" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#fetch zipcode from geocode API for each station\n", - "stations = stations + \" station, NY\"\n", - "def getzipcode(ser):\n", - " station_dict = dict()\n", - " for station in ser:\n", - " # Try to pull lat & long coordinates. If fail, print index and fetch the next station lat & long\n", - " try:\n", - " zipcode = gmaps.geocode(station)[0][\"address_components\"][-1]['long_name']\n", - " station_dict[station] = zipcode\n", - " if len(station_dict) % 50 == 0:\n", - " print(\"index =\", len(station_dict), \"zipcode =\", zipcode)\n", - " except IndexError:\n", - " print(\"index error at index=\", len(station_dict))\n", - " pass\n", - " return station_dict\n", - "station_dict = getzipcode(stations)" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "# save station & zipcode dictionary as text file\n", - "import csv\n", - "f = open(\"station_dict.txt\",\"w\")\n", - "f.write( str(station_dict) )\n", - "f.close()" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#Write Station Zipcode Dictionary to CSV file\n", - "\n", - "#Convert Dictionary to Dataframe, convert non-zipcodes to NaN, zipcodes to integers\n", - "zipcode_df = pd.DataFrame(list(station_dict.items()), columns=['STATION', 'zipcode'])\n", - "latlong_df = pd.DataFrame(list(latlong_dict.items()), columns=['STATION', 'latlong'])\n", - "zipcode_df[\"STATION\"] = zipcode_df[\"STATION\"].replace(\"\\sstation,\\sNY\",\"\", regex = True)\n", - "zipcode_df.head()" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "# Fix some of the missing/incorrect zipcodes\n", - "zipcode_df[\"zipcode\"] = pd.to_numeric(zipcode_df[\"zipcode\"],errors='coerce',downcast='integer')\n", - "\n", - "# Find assigned values (google geocode could not locate the zipcode, or wrongly identifies the zipcode)\n", - "unassigned = zipcode_df[(zipcode_df.zipcode.isnull()) | (zipcode_df.zipcode < 10000)]\n", - "unassigned" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#Assign mislocated or missing zipcodes manually\n", - "zipcode_df.iloc[0,1] = 11207.0\n", - "zipcode_df.iloc[4,1] = 10018.0\n", - "zipcode_df.iloc[8,1] = 10003.0\n", - "zipcode_df.iloc[10,1] = 10012.0\n", - "zipcode_df.iloc[16,1] = 10002.0\n", - "zipcode_df.iloc[41,1] = 11217.0\n", - "zipcode_df.iloc[61,1] = 11219.0\n", - "zipcode_df.iloc[67,1] = 10019.0\n", - "zipcode_df.iloc[78,1] = 11207.0\n", - "zipcode_df.iloc[87,1] = 11430.0\n", - "zipcode_df.iloc[107,1] = 11418.0\n", - "zipcode_df.iloc[129,1] = 10023.0\n", - "zipcode_df.iloc[152,1] = 11416.0\n", - "zipcode_df.iloc[190,1] = 11375.0\n", - "zipcode_df.iloc[193,1] = 11415.0\n", - "zipcode_df.iloc[229,1] = 11432.0\n", - "# skip new jersey stations & lackawanna\n", - "zipcode_df.iloc[245,1] = 10001.0\n", - "# skip new jersey newark city\n", - "zipcode_df.iloc[258,1] = 10040.0\n", - "zipcode_df.iloc[324,1] = 11101.0\n", - "zipcode_df.iloc[330,1] = 11377.0\n", - "zipcode_df.iloc[334,1] = 11372.0\n", - "zipcode_df.iloc[352,1] = 11212.0\n", - "#RIT-MANHATTAN\n", - "\n", - "# unassigned = zipcode_df[(zipcode_df.zipcode.isnull()) | (zipcode_df.zipcode < 10000)]\n", - "# unassigned" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#write to csv\n", - "zipcode_df.zipcode = zipcode_df.zipcode.astype(\"int64\")\n", - "#zipcode_df.info()" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#Save the zipcode dataframe to CSV file\n", - "zipcode_df.to_csv(\"zipcode_df.csv\")" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "## Get station coordinates" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "# Fetch lat & long data from goeocode API\n", - "stations = stations + \" station, NY\"\n", - "def getlatlong(ser):\n", - " latlong_dict = dict()\n", - " for station in ser:\n", - " # Try to pull lat & long coordinates. If fail, print index and fetch the next station lat & long\n", - " try:\n", - " latlong = gmaps.geocode(station)[0][\"geometry\"][\"location\"]\n", - " latlong_dict[station] = latlong\n", - " if len(latlong_dict) % 50 == 0:\n", - " print(\"index =\", len(latlong_dict), \"latlong =\", latlong)\n", - " except IndexError:\n", - " print(\"index error at index=\", len(latlong_dict))\n", - " pass\n", - " return latlong_dict\n", - "latlong_dict = getlatlong(stations)" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "# save dictionary as text\n", - "import csv\n", - "f = open(\"latlong_dict.txt\",\"w\")\n", - "f.write( str(station_dict) )\n", - "f.close()" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#Write Dictionary to CSV file\n", - "latlong_df = pd.DataFrame.from_dict(latlong_dict).T\n", - "#latlong_df[\"STATION\"] = latlong_df.index\n", - "latlong_df.reset_index(level=latlong_df.index.names, inplace=True)\n", - "latlong_df = latlong_df.rename(columns={\"index\": \"STATION\"})\n", - "latlong_df[\"STATION\"]= latlong_df[\"STATION\"].replace(\"\\sstation,\\sNY\",\"\", regex = True)\n", - "latlong_df.lng.describe()" - ] - } - ], - "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": "3.6.4" - } - }, - "nbformat": 4, - "nbformat_minor": 2 -} diff --git a/01-MTA_Turnstile_EDA/Benson_geolocation_final.ipynb b/01-MTA_Turnstile_EDA/Benson_geolocation_final.ipynb deleted file mode 100644 index 12472a2..0000000 --- a/01-MTA_Turnstile_EDA/Benson_geolocation_final.ipynb +++ /dev/null @@ -1,272 +0,0 @@ -{ - "cells": [ - { - "cell_type": "code", - "execution_count": 1, - "metadata": {}, - "outputs": [], - "source": [ - "#Initial Imports\n", - "import pandas as pd\n", - "import numpy as np\n", - "import matplotlib.pyplot as plt\n", - "import seaborn as sns\n", - "import datetime as dt\n", - "import glob\n", - "import os\n", - "import googlemaps" - ] - }, - { - "cell_type": "code", - "execution_count": 2, - "metadata": {}, - "outputs": [], - "source": [ - "# Read NYC turntable for the month of march, from march 3rd to march 30th (4weeks)\n", - "#df = pd.read_csv(\"NYCT180310.csv\")\n", - "path =r'/Users/petermin/metis/01-benson' # use your path\n", - "all_files = glob.glob(os.path.join(path, \"NYCT*.csv\"))\n", - "all_files\n", - "df = pd.concat((pd.read_csv(f) for f in all_files))\n", - "df.columns = df.columns.str.strip()\n", - "df;" - ] - }, - { - "cell_type": "code", - "execution_count": 5, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "array(['59 ST', '5 AV/59 ST', '57 ST-7 AV', '49 ST', 'TIMES SQ-42 ST',\n", - " '34 ST-HERALD SQ', '28 ST', '23 ST', '14 ST-UNION SQ', '8 ST-NYU'],\n", - " dtype=object)" - ] - }, - "execution_count": 5, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "# Get the list of station names in the dataframe\n", - "stations = df[\"STATION\"].unique()\n", - "stations[:10]" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "## Get Station Zipcode" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#fetch zipcode from geocode API for each station\n", - "stations = stations + \" station, NY\"\n", - "def getzipcode(ser):\n", - " station_dict = dict()\n", - " for station in ser:\n", - " # Try to pull lat & long coordinates. If fail, print index and fetch the next station lat & long\n", - " try:\n", - " zipcode = gmaps.geocode(station)[0][\"address_components\"][-1]['long_name']\n", - " station_dict[station] = zipcode\n", - " if len(station_dict) % 50 == 0:\n", - " print(\"index =\", len(station_dict), \"zipcode =\", zipcode)\n", - " except IndexError:\n", - " print(\"index error at index=\", len(station_dict))\n", - " pass\n", - " return station_dict\n", - "station_dict = getzipcode(stations)" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "# save station & zipcode dictionary as text file\n", - "import csv\n", - "f = open(\"station_dict.txt\",\"w\")\n", - "f.write( str(station_dict) )\n", - "f.close()" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#Write Station Zipcode Dictionary to CSV file\n", - "\n", - "#Convert Dictionary to Dataframe, convert non-zipcodes to NaN, zipcodes to integers\n", - "zipcode_df = pd.DataFrame(list(station_dict.items()), columns=['STATION', 'zipcode'])\n", - "latlong_df = pd.DataFrame(list(latlong_dict.items()), columns=['STATION', 'latlong'])\n", - "zipcode_df[\"STATION\"] = zipcode_df[\"STATION\"].replace(\"\\sstation,\\sNY\",\"\", regex = True)\n", - "zipcode_df.head()" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "# Fix some of the missing/incorrect zipcodes\n", - "zipcode_df[\"zipcode\"] = pd.to_numeric(zipcode_df[\"zipcode\"],errors='coerce',downcast='integer')\n", - "\n", - "# Find assigned values (google geocode could not locate the zipcode, or wrongly identifies the zipcode)\n", - "unassigned = zipcode_df[(zipcode_df.zipcode.isnull()) | (zipcode_df.zipcode < 10000)]\n", - "unassigned" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#Assign mislocated or missing zipcodes manually\n", - "zipcode_df.iloc[0,1] = 11207.0\n", - "zipcode_df.iloc[4,1] = 10018.0\n", - "zipcode_df.iloc[8,1] = 10003.0\n", - "zipcode_df.iloc[10,1] = 10012.0\n", - "zipcode_df.iloc[16,1] = 10002.0\n", - "zipcode_df.iloc[41,1] = 11217.0\n", - "zipcode_df.iloc[61,1] = 11219.0\n", - "zipcode_df.iloc[67,1] = 10019.0\n", - "zipcode_df.iloc[78,1] = 11207.0\n", - "zipcode_df.iloc[87,1] = 11430.0\n", - "zipcode_df.iloc[107,1] = 11418.0\n", - "zipcode_df.iloc[129,1] = 10023.0\n", - "zipcode_df.iloc[152,1] = 11416.0\n", - "zipcode_df.iloc[190,1] = 11375.0\n", - "zipcode_df.iloc[193,1] = 11415.0\n", - "zipcode_df.iloc[229,1] = 11432.0\n", - "# skip new jersey stations & lackawanna\n", - "zipcode_df.iloc[245,1] = 10001.0\n", - "# skip new jersey newark city\n", - "zipcode_df.iloc[258,1] = 10040.0\n", - "zipcode_df.iloc[324,1] = 11101.0\n", - "zipcode_df.iloc[330,1] = 11377.0\n", - "zipcode_df.iloc[334,1] = 11372.0\n", - "zipcode_df.iloc[352,1] = 11212.0\n", - "#RIT-MANHATTAN\n", - "\n", - "# unassigned = zipcode_df[(zipcode_df.zipcode.isnull()) | (zipcode_df.zipcode < 10000)]\n", - "# unassigned" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#write to csv\n", - "zipcode_df.zipcode = zipcode_df.zipcode.astype(\"int64\")\n", - "#zipcode_df.info()" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#Save the zipcode dataframe to CSV file\n", - "zipcode_df.to_csv(\"zipcode_df.csv\")" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "## Get station coordinates" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "# Fetch lat & long data from goeocode API\n", - "stations = stations + \" station, NY\"\n", - "def getlatlong(ser):\n", - " latlong_dict = dict()\n", - " for station in ser:\n", - " # Try to pull lat & long coordinates. If fail, print index and fetch the next station lat & long\n", - " try:\n", - " latlong = gmaps.geocode(station)[0][\"geometry\"][\"location\"]\n", - " latlong_dict[station] = latlong\n", - " if len(latlong_dict) % 50 == 0:\n", - " print(\"index =\", len(latlong_dict), \"latlong =\", latlong)\n", - " except IndexError:\n", - " print(\"index error at index=\", len(latlong_dict))\n", - " pass\n", - " return latlong_dict\n", - "latlong_dict = getlatlong(stations)" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "# save dictionary as text\n", - "import csv\n", - "f = open(\"latlong_dict.txt\",\"w\")\n", - "f.write( str(station_dict) )\n", - "f.close()" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#Write Dictionary to CSV file\n", - "latlong_df = pd.DataFrame.from_dict(latlong_dict).T\n", - "#latlong_df[\"STATION\"] = latlong_df.index\n", - "latlong_df.reset_index(level=latlong_df.index.names, inplace=True)\n", - "latlong_df = latlong_df.rename(columns={\"index\": \"STATION\"})\n", - "latlong_df[\"STATION\"]= latlong_df[\"STATION\"].replace(\"\\sstation,\\sNY\",\"\", regex = True)\n", - "latlong_df.lng.describe()" - ] - } - ], - "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": "3.6.4" - } - }, - "nbformat": 4, - "nbformat_minor": 2 -} diff --git a/02-Movie_Opening_Gross_Prediction/scrap/.ipynb_checkpoints/Scrapcode selenium testing-checkpoint.ipynb b/02-Movie_Opening_Gross_Prediction/scrap/.ipynb_checkpoints/Scrapcode selenium testing-checkpoint.ipynb new file mode 100644 index 0000000..edaec4b --- /dev/null +++ b/02-Movie_Opening_Gross_Prediction/scrap/.ipynb_checkpoints/Scrapcode selenium testing-checkpoint.ipynb @@ -0,0 +1,141 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# Selenium test" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "metadata": {}, + "outputs": [], + "source": [ + "# import selenium and run chromedriver\n", + "\n", + "# !pip install selenium \n", + "# download chromedriver: https://sites.google.com/a/chromium.org/chromedriver/downloads \n", + "\n", + "from selenium import webdriver\n", + "from selenium.webdriver.common.keys import Keys\n", + "from selenium.webdriver.support.ui import Select\n", + "import time\n", + "import os\n", + "\n", + "chromedriver = \"/Users/petermin/Downloads/chromedriver\" # path to the chromedriver executable\n", + "os.environ[\"webdriver.chrome.driver\"] = chromedriver\n", + "driver = webdriver.Chrome(chromedriver)\n", + "#driver.get(\"https://trends.google.com/trends/\")" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "metadata": {}, + "outputs": [ + { + "ename": "NoSuchElementException", + "evalue": "Message: no such element: Unable to locate element: {\"method\":\"css selector\",\"selector\":\"input\"}\n (Session info: chrome=70.0.3538.110)\n (Driver info: chromedriver=2.37.544337 (8c0344a12e552148c185f7d5117db1f28d6c9e85),platform=Mac OS X 10.14.1 x86_64)\n", + "output_type": "error", + "traceback": [ + "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", + "\u001b[0;31mNoSuchElementException\u001b[0m Traceback (most recent call last)", + "\u001b[0;32m\u001b[0m in \u001b[0;36m\u001b[0;34m()\u001b[0m\n\u001b[1;32m 2\u001b[0m \u001b[0;31m#link with geo set as US and youtube searches\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 3\u001b[0m \u001b[0mdriver\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'https://trends.google.com/trends/explore?geo=US&gprop=youtube'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0msearch_term\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdriver\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfind_element_by_css_selector\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"input\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 5\u001b[0m \u001b[0;31m#time.sleep(1)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 6\u001b[0m \u001b[0msearch_term\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msend_keys\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Dunkirk\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/selenium/webdriver/remote/webdriver.py\u001b[0m in \u001b[0;36mfind_element_by_css_selector\u001b[0;34m(self, css_selector)\u001b[0m\n\u001b[1;32m 587\u001b[0m \u001b[0melement\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdriver\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfind_element_by_css_selector\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'#foo'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 588\u001b[0m \"\"\"\n\u001b[0;32m--> 589\u001b[0;31m \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfind_element\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mby\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mBy\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mCSS_SELECTOR\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcss_selector\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 590\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 591\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mfind_elements_by_css_selector\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcss_selector\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/selenium/webdriver/remote/webdriver.py\u001b[0m in \u001b[0;36mfind_element\u001b[0;34m(self, by, value)\u001b[0m\n\u001b[1;32m 953\u001b[0m return self.execute(Command.FIND_ELEMENT, {\n\u001b[1;32m 954\u001b[0m \u001b[0;34m'using'\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mby\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 955\u001b[0;31m 'value': value})['value']\n\u001b[0m\u001b[1;32m 956\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 957\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mfind_elements\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mby\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mBy\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mID\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/selenium/webdriver/remote/webdriver.py\u001b[0m in \u001b[0;36mexecute\u001b[0;34m(self, driver_command, params)\u001b[0m\n\u001b[1;32m 310\u001b[0m \u001b[0mresponse\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcommand_executor\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexecute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdriver_command\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mparams\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 311\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mresponse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 312\u001b[0;31m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0merror_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcheck_response\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mresponse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 313\u001b[0m response['value'] = self._unwrap_value(\n\u001b[1;32m 314\u001b[0m response.get('value', None))\n", + "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/selenium/webdriver/remote/errorhandler.py\u001b[0m in \u001b[0;36mcheck_response\u001b[0;34m(self, response)\u001b[0m\n\u001b[1;32m 240\u001b[0m \u001b[0malert_text\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'alert'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'text'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 241\u001b[0m \u001b[0;32mraise\u001b[0m \u001b[0mexception_class\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmessage\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mscreen\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstacktrace\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malert_text\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 242\u001b[0;31m \u001b[0;32mraise\u001b[0m \u001b[0mexception_class\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmessage\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mscreen\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstacktrace\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 243\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 244\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m_value_or_default\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mobj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdefault\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;31mNoSuchElementException\u001b[0m: Message: no such element: Unable to locate element: {\"method\":\"css selector\",\"selector\":\"input\"}\n (Session info: chrome=70.0.3538.110)\n (Driver info: chromedriver=2.37.544337 (8c0344a12e552148c185f7d5117db1f28d6c9e85),platform=Mac OS X 10.14.1 x86_64)\n" + ] + } + ], + "source": [ + "# search the movie in google trends\n", + "#link with geo set as US and youtube searches\n", + "driver.get('https://trends.google.com/trends/explore?geo=US&gprop=youtube')\n", + "search_term = driver.find_element_by_css_selector(\"input\")\n", + "#time.sleep(1)\n", + "search_term.send_keys(\"Dunkirk\")\n", + "#time.sleep(2)\n", + "search_term.send_keys(Keys.RETURN)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#select_date = Select(driver.find_element_by_css_selector(\"md-select-value\"))\n", + "#select_date.click()\n", + "#driver.find_element_by_xpath(\"//md-option[@name='Custom time range...']/option[text()='option_text']\").click()" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#select_date.select_by_visible_text('select_option_30')" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#select_date = Select(driver.find_element_by_css_selector(\"md-select-value\"))\n", + "#select_date = Select(driver.find_element_by_id(\"select_value_label_16\"))\n", + "\n", + "#select = driver.find_element_by_xpath(\"//custom-date-picker/ng-include/md-select\")\n", + "#driver.find_element_by_xpath(\"//md-select-value\").click()\n", + "\n", + "#driver.find_element_by_id(\"select_container_18\")\n", + "\n", + "\n", + "# select = driver.find_element_by_xpath(\"//md-select-menu/md-content\")\n", + "# for option in select.find_elements_by_tag_name('md-option'):\n", + "# if option.id == \"select_option_30\":\n", + "# option.click()" + ] + } + ], + "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": "3.6.4" + }, + "toc": { + "base_numbering": 1, + "nav_menu": {}, + "number_sections": true, + "sideBar": true, + "skip_h1_title": false, + "title_cell": "Table of Contents", + "title_sidebar": "Contents", + "toc_cell": false, + "toc_position": {}, + "toc_section_display": true, + "toc_window_display": false + } + }, + "nbformat": 4, + "nbformat_minor": 2 +} diff --git a/02-Movie_Opening_Gross_Prediction/scrap/Scrapcode selenium testing.ipynb b/02-Movie_Opening_Gross_Prediction/scrap/Scrapcode selenium testing.ipynb index 347e503..bd43f98 100644 --- a/02-Movie_Opening_Gross_Prediction/scrap/Scrapcode selenium testing.ipynb +++ b/02-Movie_Opening_Gross_Prediction/scrap/Scrapcode selenium testing.ipynb @@ -9,7 +9,7 @@ }, { "cell_type": "code", - "execution_count": 1, + "execution_count": 4, "metadata": {}, "outputs": [], "source": [ @@ -32,15 +32,31 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 5, "metadata": {}, - "outputs": [], + "outputs": [ + { + "ename": "NoSuchElementException", + "evalue": "Message: no such element: Unable to locate element: {\"method\":\"css selector\",\"selector\":\"input\"}\n (Session info: chrome=70.0.3538.110)\n (Driver info: chromedriver=2.37.544337 (8c0344a12e552148c185f7d5117db1f28d6c9e85),platform=Mac OS X 10.14.1 x86_64)\n", + "output_type": "error", + "traceback": [ + "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", + "\u001b[0;31mNoSuchElementException\u001b[0m Traceback (most recent call last)", + "\u001b[0;32m\u001b[0m in \u001b[0;36m\u001b[0;34m()\u001b[0m\n\u001b[1;32m 2\u001b[0m \u001b[0;31m#link with geo set as US and youtube searches\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 3\u001b[0m \u001b[0mdriver\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'https://trends.google.com/trends/explore?geo=US&gprop=youtube'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0msearch_term\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdriver\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfind_element_by_css_selector\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"input\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 5\u001b[0m \u001b[0;31m#time.sleep(1)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 6\u001b[0m \u001b[0msearch_term\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msend_keys\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Dunkirk\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/selenium/webdriver/remote/webdriver.py\u001b[0m in \u001b[0;36mfind_element_by_css_selector\u001b[0;34m(self, css_selector)\u001b[0m\n\u001b[1;32m 587\u001b[0m \u001b[0melement\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdriver\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfind_element_by_css_selector\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'#foo'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 588\u001b[0m \"\"\"\n\u001b[0;32m--> 589\u001b[0;31m \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfind_element\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mby\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mBy\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mCSS_SELECTOR\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcss_selector\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 590\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 591\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mfind_elements_by_css_selector\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcss_selector\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/selenium/webdriver/remote/webdriver.py\u001b[0m in \u001b[0;36mfind_element\u001b[0;34m(self, by, value)\u001b[0m\n\u001b[1;32m 953\u001b[0m return self.execute(Command.FIND_ELEMENT, {\n\u001b[1;32m 954\u001b[0m \u001b[0;34m'using'\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mby\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 955\u001b[0;31m 'value': value})['value']\n\u001b[0m\u001b[1;32m 956\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 957\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mfind_elements\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mby\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mBy\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mID\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/selenium/webdriver/remote/webdriver.py\u001b[0m in \u001b[0;36mexecute\u001b[0;34m(self, driver_command, params)\u001b[0m\n\u001b[1;32m 310\u001b[0m \u001b[0mresponse\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcommand_executor\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexecute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdriver_command\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mparams\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 311\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mresponse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 312\u001b[0;31m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0merror_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcheck_response\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mresponse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 313\u001b[0m response['value'] = self._unwrap_value(\n\u001b[1;32m 314\u001b[0m response.get('value', None))\n", + "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/selenium/webdriver/remote/errorhandler.py\u001b[0m in \u001b[0;36mcheck_response\u001b[0;34m(self, response)\u001b[0m\n\u001b[1;32m 240\u001b[0m \u001b[0malert_text\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'alert'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'text'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 241\u001b[0m \u001b[0;32mraise\u001b[0m \u001b[0mexception_class\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmessage\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mscreen\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstacktrace\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malert_text\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 242\u001b[0;31m \u001b[0;32mraise\u001b[0m \u001b[0mexception_class\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmessage\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mscreen\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstacktrace\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 243\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 244\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m_value_or_default\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mobj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdefault\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;31mNoSuchElementException\u001b[0m: Message: no such element: Unable to locate element: {\"method\":\"css selector\",\"selector\":\"input\"}\n (Session info: chrome=70.0.3538.110)\n (Driver info: chromedriver=2.37.544337 (8c0344a12e552148c185f7d5117db1f28d6c9e85),platform=Mac OS X 10.14.1 x86_64)\n" + ] + } + ], "source": [ "# search the movie in google trends\n", "#link with geo set as US and youtube searches\n", "driver.get('https://trends.google.com/trends/explore?geo=US&gprop=youtube')\n", "search_term = driver.find_element_by_css_selector(\"input\")\n", - "#time.sleep(1)\n", + "# time.sleep(1)\n", "search_term.send_keys(\"Dunkirk\")\n", "#time.sleep(2)\n", "search_term.send_keys(Keys.RETURN)" diff --git a/04-Aspect_Based_Opinion_Mining/.DS_Store b/04-Aspect_Based_Opinion_Mining/.DS_Store index a741147c01e8de794b7d92ebebf511718c4895b3..4be7a00a8da793df433e151cf6f22b44acbf9061 100644 GIT binary patch delta 215 zcmZn(XbISmFUWX&GNW(9J@ydS`mD_A1ewt%q1KVbHgf}xScWLZ(!$;*Y+HroifF>;`ZzQL(pWpbS8 iCp@y3#ici!iJ#`!%&zc@W%7R!Q7m>6Va8-L@e=@=hdA5- diff --git a/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/01-Build_Model-checkpoint.ipynb b/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/01-Build_Model-checkpoint.ipynb index fb3ed1f..2bbde29 100644 --- a/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/01-Build_Model-checkpoint.ipynb +++ b/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/01-Build_Model-checkpoint.ipynb @@ -16,12 +16,12 @@ }, { "cell_type": "code", - "execution_count": 1, + "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "import xml.etree.ElementTree as ET\n", - "tree = ET.parse('Restaurants_Train.xml')\n", + "tree = ET.parse('../data/Restaurants_Train.xml')\n", "root = tree.getroot()" ] }, @@ -63,34 +63,6 @@ "labeled_df.head()" ] }, - { - "cell_type": "code", - "execution_count": 21, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": 55, - "metadata": {}, - "outputs": [], - "source": [] - }, { "cell_type": "markdown", "metadata": {}, @@ -345,13 +317,6 @@ "At this point, we can move on to 02-Sentiment analysis notebook, which will load the fitted Naive bayes model." ] }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, { "cell_type": "code", "execution_count": 48, @@ -3423,13 +3388,6 @@ "pred_df.head()" ] }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, { "cell_type": "markdown", "metadata": {}, @@ -3531,7 +3489,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.6.4" + "version": "3.6.8" }, "toc": { "base_numbering": 1, diff --git a/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/02-Sentiment_Analysis-checkpoint.ipynb b/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/02-Sentiment_Analysis-checkpoint.ipynb index 9ebd3b4..892ceab 100644 --- a/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/02-Sentiment_Analysis-checkpoint.ipynb +++ b/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/02-Sentiment_Analysis-checkpoint.ipynb @@ -127,6 +127,10 @@ "outputs": [], "source": [ "def check_similarity(aspects, word):\n", + " '''\n", + " checks for word2vec similarity values between category word and the term\n", + " returns most similar word\n", + " '''\n", " similarity = []\n", " for aspect in aspects:\n", " similarity.append(word2vec.n_similarity([aspect], [word]))\n", @@ -326,27 +330,6 @@ "# displacy.render(spacy(sentence), style='dep',jupyter=True)" ] }, - { - "cell_type": "code", - "execution_count": 148, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "\"he's a\"" - ] - }, - "execution_count": 148, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "# test code for special char function\n", - "# remove_special_char(\"he's/a\")" - ] - }, { "cell_type": "code", "execution_count": 59, @@ -364,40 +347,11 @@ "metadata": {}, "outputs": [], "source": [ - "# test code\n", + "# test case 1\n", "terms_dict={'ambience':Counter(), 'food':Counter(), 'price':Counter(), 'service':Counter(),'misc':Counter()}\n", "aspect_sent={'ambience':Counter(), 'food':Counter(), 'price':Counter(), 'service':Counter(),'misc':Counter()}\n", "review = \"Our waiter was not very helpful, and the music was terrible.\"\n", - "test1, test2 = review_pipe(review, aspect_sent, terms_dict)" - ] - }, - { - "cell_type": "code", - "execution_count": 17, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "({'ambience': Counter({'neg': 1}),\n", - " 'food': Counter(),\n", - " 'price': Counter(),\n", - " 'service': Counter({'neg': 1}),\n", - " 'misc': Counter()},\n", - " {'ambience': Counter({'music': -1}),\n", - " 'food': Counter(),\n", - " 'price': Counter(),\n", - " 'service': Counter({'waiter': -1}),\n", - " 'misc': Counter()})" - ] - }, - "execution_count": 17, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "test1, test2" + "review_pipe(review, aspect_sent, terms_dict)" ] }, { @@ -426,6 +380,7 @@ } ], "source": [ + "# test case 2\n", "test1={'ambience':Counter(), 'food':Counter(), 'price':Counter(), 'service':Counter(),'misc':Counter()}\n", "test2={'ambience':Counter(), 'food':Counter(), 'price':Counter(), 'service':Counter(),'misc':Counter()}\n", "review = \"top notch\"\n", @@ -443,13 +398,6 @@ "3) update aspect dictionary" ] }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, { "cell_type": "code", "execution_count": 82, @@ -474,7 +422,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "# Load up dataframe and test" + "## Load up dataframe and test" ] }, { @@ -1736,27 +1684,8 @@ } ], "source": [ - "pitt[(pitt.name!='\"Meat & Potatoes\"') & (pitt.categories.str.contains(\";American\"))].sort_values(\"review_count\",ascending=False)" - ] - }, - { - "cell_type": "code", - "execution_count": 53, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "1101" - ] - }, - "execution_count": 53, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "len(primanti)" + "pitt[(pitt.name!='\"Meat & Potatoes\"') & (pitt.categories.str.contains(\";American\"))].sort_values(\"review_count\",ascending=False)\n", + "pitt.head()" ] }, { @@ -3653,20 +3582,6 @@ "source": [ "Now that we have pickled files, we can generate visualizations using plotly in the 3rd notebook 03." ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] } ], "metadata": { @@ -3685,7 +3600,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.6.4" + "version": "3.6.8" }, "toc": { "base_numbering": 1, diff --git a/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/03 - Scrap_code_Plotly-checkpoint.ipynb b/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/03 - Plotly-checkpoint.ipynb similarity index 99% rename from 04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/03 - Scrap_code_Plotly-checkpoint.ipynb rename to 04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/03 - Plotly-checkpoint.ipynb index bdd8449..1db5311 100644 --- a/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/03 - Scrap_code_Plotly-checkpoint.ipynb +++ b/04-Aspect_Based_Opinion_Mining/code/.ipynb_checkpoints/03 - Plotly-checkpoint.ipynb @@ -290,7 +290,7 @@ ], "source": [ "import plotly\n", - "# ***fill in username and API Key for plotly here***\n", + "# *fill in username and API Key for plotly here*\n", "plotly.tools.set_credentials_file(username='####', api_key='######')\n", "import plotly.plotly as py\n", "import plotly.graph_objs as go\n", @@ -420,13 +420,6 @@ "\n", "py.iplot([trace])" ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] } ], "metadata": { @@ -445,7 +438,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.6.4" + "version": "3.6.8" }, "toc": { "base_numbering": 1, diff --git a/04-Aspect_Based_Opinion_Mining/code/01-Build_Model.ipynb b/04-Aspect_Based_Opinion_Mining/code/01-Build_Model.ipynb index d05d17b..2bbde29 100644 --- a/04-Aspect_Based_Opinion_Mining/code/01-Build_Model.ipynb +++ b/04-Aspect_Based_Opinion_Mining/code/01-Build_Model.ipynb @@ -25,26 +25,6 @@ "root = tree.getroot()" ] }, - { - "cell_type": "code", - "execution_count": 3, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "" - ] - }, - "execution_count": 3, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "root" - ] - }, { "cell_type": "code", "execution_count": 6, @@ -83,34 +63,6 @@ "labeled_df.head()" ] }, - { - "cell_type": "code", - "execution_count": 21, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": 55, - "metadata": {}, - "outputs": [], - "source": [] - }, { "cell_type": "markdown", "metadata": {}, @@ -365,13 +317,6 @@ "At this point, we can move on to 02-Sentiment analysis notebook, which will load the fitted Naive bayes model." ] }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, { "cell_type": "code", "execution_count": 48, @@ -3443,13 +3388,6 @@ "pred_df.head()" ] }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, { "cell_type": "markdown", "metadata": {}, @@ -3551,7 +3489,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.6.4" + "version": "3.6.8" }, "toc": { "base_numbering": 1, diff --git a/04-Aspect_Based_Opinion_Mining/code/02-Sentiment_Analysis.ipynb b/04-Aspect_Based_Opinion_Mining/code/02-Sentiment_Analysis.ipynb index 9ebd3b4..892ceab 100644 --- a/04-Aspect_Based_Opinion_Mining/code/02-Sentiment_Analysis.ipynb +++ b/04-Aspect_Based_Opinion_Mining/code/02-Sentiment_Analysis.ipynb @@ -127,6 +127,10 @@ "outputs": [], "source": [ "def check_similarity(aspects, word):\n", + " '''\n", + " checks for word2vec similarity values between category word and the term\n", + " returns most similar word\n", + " '''\n", " similarity = []\n", " for aspect in aspects:\n", " similarity.append(word2vec.n_similarity([aspect], [word]))\n", @@ -326,27 +330,6 @@ "# displacy.render(spacy(sentence), style='dep',jupyter=True)" ] }, - { - "cell_type": "code", - "execution_count": 148, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "\"he's a\"" - ] - }, - "execution_count": 148, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "# test code for special char function\n", - "# remove_special_char(\"he's/a\")" - ] - }, { "cell_type": "code", "execution_count": 59, @@ -364,40 +347,11 @@ "metadata": {}, "outputs": [], "source": [ - "# test code\n", + "# test case 1\n", "terms_dict={'ambience':Counter(), 'food':Counter(), 'price':Counter(), 'service':Counter(),'misc':Counter()}\n", "aspect_sent={'ambience':Counter(), 'food':Counter(), 'price':Counter(), 'service':Counter(),'misc':Counter()}\n", "review = \"Our waiter was not very helpful, and the music was terrible.\"\n", - "test1, test2 = review_pipe(review, aspect_sent, terms_dict)" - ] - }, - { - "cell_type": "code", - "execution_count": 17, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "({'ambience': Counter({'neg': 1}),\n", - " 'food': Counter(),\n", - " 'price': Counter(),\n", - " 'service': Counter({'neg': 1}),\n", - " 'misc': Counter()},\n", - " {'ambience': Counter({'music': -1}),\n", - " 'food': Counter(),\n", - " 'price': Counter(),\n", - " 'service': Counter({'waiter': -1}),\n", - " 'misc': Counter()})" - ] - }, - "execution_count": 17, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "test1, test2" + "review_pipe(review, aspect_sent, terms_dict)" ] }, { @@ -426,6 +380,7 @@ } ], "source": [ + "# test case 2\n", "test1={'ambience':Counter(), 'food':Counter(), 'price':Counter(), 'service':Counter(),'misc':Counter()}\n", "test2={'ambience':Counter(), 'food':Counter(), 'price':Counter(), 'service':Counter(),'misc':Counter()}\n", "review = \"top notch\"\n", @@ -443,13 +398,6 @@ "3) update aspect dictionary" ] }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, { "cell_type": "code", "execution_count": 82, @@ -474,7 +422,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "# Load up dataframe and test" + "## Load up dataframe and test" ] }, { @@ -1736,27 +1684,8 @@ } ], "source": [ - "pitt[(pitt.name!='\"Meat & Potatoes\"') & (pitt.categories.str.contains(\";American\"))].sort_values(\"review_count\",ascending=False)" - ] - }, - { - "cell_type": "code", - "execution_count": 53, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "1101" - ] - }, - "execution_count": 53, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "len(primanti)" + "pitt[(pitt.name!='\"Meat & Potatoes\"') & (pitt.categories.str.contains(\";American\"))].sort_values(\"review_count\",ascending=False)\n", + "pitt.head()" ] }, { @@ -3653,20 +3582,6 @@ "source": [ "Now that we have pickled files, we can generate visualizations using plotly in the 3rd notebook 03." ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] } ], "metadata": { @@ -3685,7 +3600,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.6.4" + "version": "3.6.8" }, "toc": { "base_numbering": 1, diff --git a/04-Aspect_Based_Opinion_Mining/code/03 - Scrap_code_Plotly.ipynb b/04-Aspect_Based_Opinion_Mining/code/03 - Plotly.ipynb similarity index 99% rename from 04-Aspect_Based_Opinion_Mining/code/03 - Scrap_code_Plotly.ipynb rename to 04-Aspect_Based_Opinion_Mining/code/03 - Plotly.ipynb index bdd8449..1db5311 100644 --- a/04-Aspect_Based_Opinion_Mining/code/03 - Scrap_code_Plotly.ipynb +++ b/04-Aspect_Based_Opinion_Mining/code/03 - Plotly.ipynb @@ -290,7 +290,7 @@ ], "source": [ "import plotly\n", - "# ***fill in username and API Key for plotly here***\n", + "# *fill in username and API Key for plotly here*\n", "plotly.tools.set_credentials_file(username='####', api_key='######')\n", "import plotly.plotly as py\n", "import plotly.graph_objs as go\n", @@ -420,13 +420,6 @@ "\n", "py.iplot([trace])" ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] } ], "metadata": { @@ -445,7 +438,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.6.4" + "version": "3.6.8" }, "toc": { "base_numbering": 1, diff --git a/05-Uber_demand_forecasting/.DS_Store b/05-Uber_demand_forecasting/.DS_Store index 23cb3d0b73ccc2c4bddf3dd8d82f5dfecd49a817..9a3cddc75076adce70025818369c189ac6821e10 100644 GIT binary patch delta 401 zcmZn(XbIS`UXb0;*jz`!&}8xh5vR>R1U(tqkocX#znGBtlTV9@G47lENK8pXOjKM# zR8mS(CSE`|DKR-ay(qslFU2`OC%?!kr!+MpGdVvmII}8svZ?s4`gj5U;*9)qPoP*p zQEFLcYI$^eQD#bTL1J=dUb?dV2L>=ub_jqmG$0HC24x6?fkD|J62jyVVB`z|Y3EdS z@P^8AFlaM4GI%irGh{H7GxRacVpzklgW)K{ErvG?zZk7q7{wTs7_}Kq7;PEt89f*S s8G{%@7(*GO7^5KuFfl?6l7-SxiZNyLREc!<&Fl&_Y?OzC9R6?s0LM>d5&!@I delta 373 zcmZn(XbIS`UXa7c(n3eU(8yx)LlMW#KLkA)C*KnB*xV`niwVS-d|FJ5ar5LyVoHXh zqGIBrLXuL_@dCn0iOJdNMfs(9DbD#h`9)4SrK#}({KXmh<(|p;c>zVKWtpkvlXb** z)kkCk1%flHQlrz0GE;&J5|cCY(v|H$Fn|H4vV#VMCBUEzVK6Xo2rzPnfK)3x1VF@; z9lW8E91PkFjts#JF${$a4Ghy67Bg&RILL60;W5KkMrKA{MlnVuMr}qNMpH&}Mh8Ym tMs;UK7e-IUAc%G*MyN3|P#Q`xMsHpqk\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
Alphabet CityAstoriaBattery Park CityBay RidgeBaysideBedfordBelmontBensonhurst WestBloomingdaleBoerum Hill...Williamsburg (North Side)Williamsburg (South Side)Windsor TerraceWoodhavenWoodsideWorld Trade CenterYorkville EastYorkville Westtempprecip
date
2015-01-01 01:00:0056.056.029.018.03.048.03.05.012.052.0...184.0100.015.05.02.018.039.083.00.2988510.0
2015-01-01 02:00:0046.034.027.021.03.049.03.06.03.029.0...183.085.07.09.04.08.027.051.00.2988510.0
2015-01-01 03:00:0044.056.016.031.010.087.00.05.03.033.0...266.0114.09.012.010.011.020.042.00.2873560.0
2015-01-01 04:00:0041.047.014.026.05.072.01.012.05.025.0...204.0122.08.07.011.015.019.028.00.2873560.0
2015-01-01 05:00:0027.032.09.07.03.023.01.03.02.011.0...66.072.02.06.06.05.07.013.00.2873560.0
2015-01-01 06:00:0011.017.05.01.01.023.02.01.02.011.0...26.070.02.01.04.06.06.09.00.2873560.0
2015-01-01 07:00:005.011.04.02.00.03.00.03.03.05.0...13.057.02.01.03.03.03.08.00.2873560.0
2015-01-01 08:00:005.06.02.02.00.05.01.01.01.05.0...10.031.03.02.00.03.03.02.00.2873560.0
2015-01-01 09:00:003.04.02.04.00.05.00.01.01.01.0...12.027.02.01.05.02.05.07.00.2988510.0
2015-01-01 10:00:005.010.04.02.00.04.02.01.00.03.0...13.016.00.03.01.03.08.05.00.3218390.0
2015-01-01 11:00:0013.08.05.04.00.06.01.02.02.012.0...10.011.03.03.01.01.011.014.00.3333330.0
2015-01-01 12:00:0016.015.09.05.00.06.01.01.04.07.0...11.024.06.01.01.08.011.014.00.3563220.0
2015-01-01 13:00:009.013.014.01.02.09.03.03.02.04.0...23.013.02.01.07.06.08.017.00.3793100.0
2015-01-01 14:00:0015.019.016.06.01.09.01.01.06.08.0...31.018.01.01.08.08.09.019.00.4022990.0
2015-01-01 15:00:0016.018.015.05.01.016.02.02.06.07.0...32.024.02.04.01.012.023.022.00.4137930.0
2015-01-01 16:00:0016.020.022.04.01.013.01.01.03.010.0...43.024.09.03.00.017.012.024.00.4252870.0
2015-01-01 17:00:0021.015.035.06.01.017.01.01.03.012.0...37.025.07.03.05.018.014.030.00.4252870.0
2015-01-01 18:00:0015.011.027.04.02.012.00.03.03.010.0...38.026.07.02.02.022.013.037.00.4252870.0
2015-01-01 19:00:0018.021.017.06.04.021.00.00.07.016.0...44.021.03.00.03.018.013.019.00.4252870.0
2015-01-01 20:00:0017.016.014.05.02.016.00.02.05.010.0...30.027.03.00.02.08.022.026.00.4137930.0
2015-01-01 21:00:0017.018.017.08.01.016.02.01.03.012.0...43.023.05.00.02.08.019.019.00.4137930.0
2015-01-01 22:00:0016.013.013.07.02.029.01.02.03.013.0...38.029.02.02.00.010.011.020.00.4137930.0
2015-01-01 23:00:0013.016.05.06.01.017.02.00.03.020.0...39.027.01.01.04.03.09.024.00.4137930.0
2015-01-02 00:00:0013.017.08.08.01.017.00.00.03.06.0...27.019.00.03.02.00.04.013.00.4022990.0
2015-01-02 01:00:0011.010.04.04.00.011.00.00.02.012.0...31.016.00.01.02.02.04.014.00.4137930.0
2015-01-02 02:00:003.09.00.02.01.05.00.00.01.04.0...15.09.02.01.01.03.03.06.00.4022990.0
2015-01-02 03:00:003.02.01.00.00.02.00.00.01.04.0...30.09.00.00.00.02.02.05.00.4022990.0
2015-01-02 04:00:002.07.01.01.01.06.01.01.01.04.0...28.012.02.01.01.01.01.06.00.4022990.0
2015-01-02 05:00:003.04.03.03.00.04.00.00.01.00.0...13.06.01.00.01.01.09.07.00.3908050.0
2015-01-02 06:00:007.03.06.03.01.08.00.00.02.02.0...9.04.04.01.00.01.05.09.00.3908050.0
..................................................................
2015-06-29 18:00:0015.014.063.04.02.013.05.02.02.014.0...43.036.03.00.07.030.036.019.00.8390800.0
2015-06-29 19:00:0019.011.067.06.02.015.02.03.08.025.0...52.035.010.02.02.041.025.030.00.8390800.0
2015-06-29 20:00:0012.028.058.06.01.021.03.04.07.019.0...50.041.03.05.04.034.020.028.00.8275860.0
2015-06-29 21:00:0019.020.056.07.05.016.02.04.08.026.0...54.045.05.04.06.028.018.027.00.8160920.0
2015-06-29 22:00:0023.018.061.08.02.021.04.03.09.017.0...49.049.05.02.010.017.013.035.00.8045980.0
2015-06-29 23:00:0019.027.052.013.01.035.01.04.07.016.0...52.050.02.03.06.025.018.016.00.8045980.0
2015-06-30 00:00:0013.013.026.06.01.022.03.05.04.018.0...44.034.04.01.05.015.07.016.00.8045980.0
2015-06-30 01:00:007.014.011.06.01.015.01.01.03.013.0...34.017.03.00.00.010.05.07.00.8045980.0
2015-06-30 02:00:004.013.06.00.01.014.00.03.01.09.0...12.012.02.00.01.07.04.03.00.7931030.0
2015-06-30 03:00:002.08.08.00.00.05.00.01.03.00.0...9.09.00.01.04.02.03.01.00.7816090.0
2015-06-30 04:00:004.012.012.02.00.05.00.01.05.05.0...11.011.00.00.02.08.04.08.00.7701150.0
2015-06-30 05:00:009.014.014.01.01.06.00.00.05.06.0...15.010.01.03.01.07.09.013.00.7586210.0
2015-06-30 06:00:0011.018.016.05.03.016.01.02.06.016.0...26.015.04.04.05.010.018.018.00.7816090.0
2015-06-30 07:00:0029.033.025.013.02.011.02.04.05.017.0...38.032.06.03.06.013.034.040.00.8160920.0
2015-06-30 08:00:0051.024.034.013.02.021.01.03.010.028.0...50.037.010.06.08.020.068.063.00.8390800.0
2015-06-30 09:00:0042.027.049.013.03.029.02.01.07.022.0...58.035.07.05.06.016.084.071.00.8620690.0
2015-06-30 10:00:0032.025.040.012.00.032.04.01.010.024.0...44.037.05.02.07.014.064.051.00.8735630.0
2015-06-30 11:00:0018.016.027.07.01.022.01.05.04.016.0...36.032.05.01.02.021.038.031.00.8850570.0
2015-06-30 12:00:0019.025.033.06.03.021.00.00.06.019.0...33.025.04.05.03.015.028.028.00.8850570.0
2015-06-30 13:00:0022.012.047.07.00.013.00.01.04.021.0...30.013.05.03.05.016.022.021.00.9080460.0
2015-06-30 14:00:0020.015.029.09.04.022.03.02.06.06.0...28.025.07.02.08.024.014.026.00.8965520.0
2015-06-30 15:00:0025.022.038.09.01.012.03.04.03.012.0...34.020.04.02.04.016.013.028.00.8965520.0
2015-06-30 16:00:0019.015.060.010.05.026.01.01.02.012.0...34.025.04.04.07.032.023.028.00.8505750.0
2015-06-30 17:00:0023.014.062.013.03.016.03.03.07.021.0...53.026.04.03.04.039.025.030.00.8505750.0
2015-06-30 18:00:0014.014.0101.010.04.024.01.01.06.022.0...43.041.05.08.03.039.042.036.00.8390800.0
2015-06-30 19:00:0030.024.091.012.05.018.04.02.06.025.0...72.049.08.04.010.043.039.041.00.8352490.0
2015-06-30 20:00:0020.042.077.07.05.018.02.06.03.028.0...64.061.010.04.06.042.034.034.00.8237550.0
2015-06-30 21:00:0031.038.064.010.05.032.02.03.011.022.0...77.050.04.02.06.024.024.039.00.8275860.0
2015-06-30 22:00:0033.029.087.014.05.025.01.03.05.031.0...89.058.06.02.09.034.025.038.00.8275860.0
2015-06-30 23:00:0024.033.042.014.04.033.02.03.04.036.0...105.061.05.01.012.016.014.034.00.8275860.0
\n", + "

4343 rows × 142 columns

\n", + "" + ], + "text/plain": [ + " Alphabet City Astoria Battery Park City Bay Ridge \\\n", + "date \n", + "2015-01-01 01:00:00 56.0 56.0 29.0 18.0 \n", + "2015-01-01 02:00:00 46.0 34.0 27.0 21.0 \n", + "2015-01-01 03:00:00 44.0 56.0 16.0 31.0 \n", + "2015-01-01 04:00:00 41.0 47.0 14.0 26.0 \n", + "2015-01-01 05:00:00 27.0 32.0 9.0 7.0 \n", + "2015-01-01 06:00:00 11.0 17.0 5.0 1.0 \n", + "2015-01-01 07:00:00 5.0 11.0 4.0 2.0 \n", + "2015-01-01 08:00:00 5.0 6.0 2.0 2.0 \n", + "2015-01-01 09:00:00 3.0 4.0 2.0 4.0 \n", + "2015-01-01 10:00:00 5.0 10.0 4.0 2.0 \n", + "2015-01-01 11:00:00 13.0 8.0 5.0 4.0 \n", + "2015-01-01 12:00:00 16.0 15.0 9.0 5.0 \n", + "2015-01-01 13:00:00 9.0 13.0 14.0 1.0 \n", + "2015-01-01 14:00:00 15.0 19.0 16.0 6.0 \n", + "2015-01-01 15:00:00 16.0 18.0 15.0 5.0 \n", + "2015-01-01 16:00:00 16.0 20.0 22.0 4.0 \n", + "2015-01-01 17:00:00 21.0 15.0 35.0 6.0 \n", + "2015-01-01 18:00:00 15.0 11.0 27.0 4.0 \n", + "2015-01-01 19:00:00 18.0 21.0 17.0 6.0 \n", + "2015-01-01 20:00:00 17.0 16.0 14.0 5.0 \n", + "2015-01-01 21:00:00 17.0 18.0 17.0 8.0 \n", + "2015-01-01 22:00:00 16.0 13.0 13.0 7.0 \n", + "2015-01-01 23:00:00 13.0 16.0 5.0 6.0 \n", + "2015-01-02 00:00:00 13.0 17.0 8.0 8.0 \n", + "2015-01-02 01:00:00 11.0 10.0 4.0 4.0 \n", + "2015-01-02 02:00:00 3.0 9.0 0.0 2.0 \n", + "2015-01-02 03:00:00 3.0 2.0 1.0 0.0 \n", + "2015-01-02 04:00:00 2.0 7.0 1.0 1.0 \n", + "2015-01-02 05:00:00 3.0 4.0 3.0 3.0 \n", + "2015-01-02 06:00:00 7.0 3.0 6.0 3.0 \n", + "... ... ... ... ... \n", + "2015-06-29 18:00:00 15.0 14.0 63.0 4.0 \n", + "2015-06-29 19:00:00 19.0 11.0 67.0 6.0 \n", + "2015-06-29 20:00:00 12.0 28.0 58.0 6.0 \n", + "2015-06-29 21:00:00 19.0 20.0 56.0 7.0 \n", + "2015-06-29 22:00:00 23.0 18.0 61.0 8.0 \n", + "2015-06-29 23:00:00 19.0 27.0 52.0 13.0 \n", + "2015-06-30 00:00:00 13.0 13.0 26.0 6.0 \n", + "2015-06-30 01:00:00 7.0 14.0 11.0 6.0 \n", + "2015-06-30 02:00:00 4.0 13.0 6.0 0.0 \n", + "2015-06-30 03:00:00 2.0 8.0 8.0 0.0 \n", + "2015-06-30 04:00:00 4.0 12.0 12.0 2.0 \n", + "2015-06-30 05:00:00 9.0 14.0 14.0 1.0 \n", + "2015-06-30 06:00:00 11.0 18.0 16.0 5.0 \n", + "2015-06-30 07:00:00 29.0 33.0 25.0 13.0 \n", + "2015-06-30 08:00:00 51.0 24.0 34.0 13.0 \n", + "2015-06-30 09:00:00 42.0 27.0 49.0 13.0 \n", + "2015-06-30 10:00:00 32.0 25.0 40.0 12.0 \n", + "2015-06-30 11:00:00 18.0 16.0 27.0 7.0 \n", + "2015-06-30 12:00:00 19.0 25.0 33.0 6.0 \n", + "2015-06-30 13:00:00 22.0 12.0 47.0 7.0 \n", + "2015-06-30 14:00:00 20.0 15.0 29.0 9.0 \n", + "2015-06-30 15:00:00 25.0 22.0 38.0 9.0 \n", + "2015-06-30 16:00:00 19.0 15.0 60.0 10.0 \n", + "2015-06-30 17:00:00 23.0 14.0 62.0 13.0 \n", + "2015-06-30 18:00:00 14.0 14.0 101.0 10.0 \n", + "2015-06-30 19:00:00 30.0 24.0 91.0 12.0 \n", + "2015-06-30 20:00:00 20.0 42.0 77.0 7.0 \n", + "2015-06-30 21:00:00 31.0 38.0 64.0 10.0 \n", + "2015-06-30 22:00:00 33.0 29.0 87.0 14.0 \n", + "2015-06-30 23:00:00 24.0 33.0 42.0 14.0 \n", + "\n", + " Bayside Bedford Belmont Bensonhurst West \\\n", + "date \n", + "2015-01-01 01:00:00 3.0 48.0 3.0 5.0 \n", + "2015-01-01 02:00:00 3.0 49.0 3.0 6.0 \n", + "2015-01-01 03:00:00 10.0 87.0 0.0 5.0 \n", + "2015-01-01 04:00:00 5.0 72.0 1.0 12.0 \n", + "2015-01-01 05:00:00 3.0 23.0 1.0 3.0 \n", + "2015-01-01 06:00:00 1.0 23.0 2.0 1.0 \n", + "2015-01-01 07:00:00 0.0 3.0 0.0 3.0 \n", + "2015-01-01 08:00:00 0.0 5.0 1.0 1.0 \n", + "2015-01-01 09:00:00 0.0 5.0 0.0 1.0 \n", + "2015-01-01 10:00:00 0.0 4.0 2.0 1.0 \n", + "2015-01-01 11:00:00 0.0 6.0 1.0 2.0 \n", + "2015-01-01 12:00:00 0.0 6.0 1.0 1.0 \n", + "2015-01-01 13:00:00 2.0 9.0 3.0 3.0 \n", + "2015-01-01 14:00:00 1.0 9.0 1.0 1.0 \n", + "2015-01-01 15:00:00 1.0 16.0 2.0 2.0 \n", + "2015-01-01 16:00:00 1.0 13.0 1.0 1.0 \n", + "2015-01-01 17:00:00 1.0 17.0 1.0 1.0 \n", + "2015-01-01 18:00:00 2.0 12.0 0.0 3.0 \n", + "2015-01-01 19:00:00 4.0 21.0 0.0 0.0 \n", + "2015-01-01 20:00:00 2.0 16.0 0.0 2.0 \n", + "2015-01-01 21:00:00 1.0 16.0 2.0 1.0 \n", + "2015-01-01 22:00:00 2.0 29.0 1.0 2.0 \n", + "2015-01-01 23:00:00 1.0 17.0 2.0 0.0 \n", + "2015-01-02 00:00:00 1.0 17.0 0.0 0.0 \n", + "2015-01-02 01:00:00 0.0 11.0 0.0 0.0 \n", + "2015-01-02 02:00:00 1.0 5.0 0.0 0.0 \n", + "2015-01-02 03:00:00 0.0 2.0 0.0 0.0 \n", + "2015-01-02 04:00:00 1.0 6.0 1.0 1.0 \n", + "2015-01-02 05:00:00 0.0 4.0 0.0 0.0 \n", + "2015-01-02 06:00:00 1.0 8.0 0.0 0.0 \n", + "... ... ... ... ... \n", + "2015-06-29 18:00:00 2.0 13.0 5.0 2.0 \n", + "2015-06-29 19:00:00 2.0 15.0 2.0 3.0 \n", + "2015-06-29 20:00:00 1.0 21.0 3.0 4.0 \n", + "2015-06-29 21:00:00 5.0 16.0 2.0 4.0 \n", + "2015-06-29 22:00:00 2.0 21.0 4.0 3.0 \n", + "2015-06-29 23:00:00 1.0 35.0 1.0 4.0 \n", + "2015-06-30 00:00:00 1.0 22.0 3.0 5.0 \n", + "2015-06-30 01:00:00 1.0 15.0 1.0 1.0 \n", + "2015-06-30 02:00:00 1.0 14.0 0.0 3.0 \n", + "2015-06-30 03:00:00 0.0 5.0 0.0 1.0 \n", + "2015-06-30 04:00:00 0.0 5.0 0.0 1.0 \n", + "2015-06-30 05:00:00 1.0 6.0 0.0 0.0 \n", + "2015-06-30 06:00:00 3.0 16.0 1.0 2.0 \n", + "2015-06-30 07:00:00 2.0 11.0 2.0 4.0 \n", + "2015-06-30 08:00:00 2.0 21.0 1.0 3.0 \n", + "2015-06-30 09:00:00 3.0 29.0 2.0 1.0 \n", + "2015-06-30 10:00:00 0.0 32.0 4.0 1.0 \n", + "2015-06-30 11:00:00 1.0 22.0 1.0 5.0 \n", + "2015-06-30 12:00:00 3.0 21.0 0.0 0.0 \n", + "2015-06-30 13:00:00 0.0 13.0 0.0 1.0 \n", + "2015-06-30 14:00:00 4.0 22.0 3.0 2.0 \n", + "2015-06-30 15:00:00 1.0 12.0 3.0 4.0 \n", + "2015-06-30 16:00:00 5.0 26.0 1.0 1.0 \n", + "2015-06-30 17:00:00 3.0 16.0 3.0 3.0 \n", + "2015-06-30 18:00:00 4.0 24.0 1.0 1.0 \n", + "2015-06-30 19:00:00 5.0 18.0 4.0 2.0 \n", + "2015-06-30 20:00:00 5.0 18.0 2.0 6.0 \n", + "2015-06-30 21:00:00 5.0 32.0 2.0 3.0 \n", + "2015-06-30 22:00:00 5.0 25.0 1.0 3.0 \n", + "2015-06-30 23:00:00 4.0 33.0 2.0 3.0 \n", + "\n", + " Bloomingdale Boerum Hill ... \\\n", + "date ... \n", + "2015-01-01 01:00:00 12.0 52.0 ... \n", + "2015-01-01 02:00:00 3.0 29.0 ... \n", + "2015-01-01 03:00:00 3.0 33.0 ... \n", + "2015-01-01 04:00:00 5.0 25.0 ... \n", + "2015-01-01 05:00:00 2.0 11.0 ... \n", + "2015-01-01 06:00:00 2.0 11.0 ... \n", + "2015-01-01 07:00:00 3.0 5.0 ... \n", + "2015-01-01 08:00:00 1.0 5.0 ... \n", + "2015-01-01 09:00:00 1.0 1.0 ... \n", + "2015-01-01 10:00:00 0.0 3.0 ... \n", + "2015-01-01 11:00:00 2.0 12.0 ... \n", + "2015-01-01 12:00:00 4.0 7.0 ... \n", + "2015-01-01 13:00:00 2.0 4.0 ... \n", + "2015-01-01 14:00:00 6.0 8.0 ... \n", + "2015-01-01 15:00:00 6.0 7.0 ... \n", + "2015-01-01 16:00:00 3.0 10.0 ... \n", + "2015-01-01 17:00:00 3.0 12.0 ... \n", + "2015-01-01 18:00:00 3.0 10.0 ... \n", + "2015-01-01 19:00:00 7.0 16.0 ... \n", + "2015-01-01 20:00:00 5.0 10.0 ... \n", + "2015-01-01 21:00:00 3.0 12.0 ... \n", + "2015-01-01 22:00:00 3.0 13.0 ... \n", + "2015-01-01 23:00:00 3.0 20.0 ... \n", + "2015-01-02 00:00:00 3.0 6.0 ... \n", + "2015-01-02 01:00:00 2.0 12.0 ... \n", + "2015-01-02 02:00:00 1.0 4.0 ... \n", + "2015-01-02 03:00:00 1.0 4.0 ... \n", + "2015-01-02 04:00:00 1.0 4.0 ... \n", + "2015-01-02 05:00:00 1.0 0.0 ... \n", + "2015-01-02 06:00:00 2.0 2.0 ... \n", + "... ... ... ... \n", + "2015-06-29 18:00:00 2.0 14.0 ... \n", + "2015-06-29 19:00:00 8.0 25.0 ... \n", + "2015-06-29 20:00:00 7.0 19.0 ... \n", + "2015-06-29 21:00:00 8.0 26.0 ... \n", + "2015-06-29 22:00:00 9.0 17.0 ... \n", + "2015-06-29 23:00:00 7.0 16.0 ... \n", + "2015-06-30 00:00:00 4.0 18.0 ... \n", + "2015-06-30 01:00:00 3.0 13.0 ... \n", + "2015-06-30 02:00:00 1.0 9.0 ... \n", + "2015-06-30 03:00:00 3.0 0.0 ... \n", + "2015-06-30 04:00:00 5.0 5.0 ... \n", + "2015-06-30 05:00:00 5.0 6.0 ... \n", + "2015-06-30 06:00:00 6.0 16.0 ... \n", + "2015-06-30 07:00:00 5.0 17.0 ... \n", + "2015-06-30 08:00:00 10.0 28.0 ... \n", + "2015-06-30 09:00:00 7.0 22.0 ... \n", + "2015-06-30 10:00:00 10.0 24.0 ... \n", + "2015-06-30 11:00:00 4.0 16.0 ... \n", + "2015-06-30 12:00:00 6.0 19.0 ... \n", + "2015-06-30 13:00:00 4.0 21.0 ... \n", + "2015-06-30 14:00:00 6.0 6.0 ... \n", + "2015-06-30 15:00:00 3.0 12.0 ... \n", + "2015-06-30 16:00:00 2.0 12.0 ... \n", + "2015-06-30 17:00:00 7.0 21.0 ... \n", + "2015-06-30 18:00:00 6.0 22.0 ... \n", + "2015-06-30 19:00:00 6.0 25.0 ... \n", + "2015-06-30 20:00:00 3.0 28.0 ... \n", + "2015-06-30 21:00:00 11.0 22.0 ... \n", + "2015-06-30 22:00:00 5.0 31.0 ... \n", + "2015-06-30 23:00:00 4.0 36.0 ... \n", + "\n", + " Williamsburg (North Side) Williamsburg (South Side) \\\n", + "date \n", + "2015-01-01 01:00:00 184.0 100.0 \n", + "2015-01-01 02:00:00 183.0 85.0 \n", + "2015-01-01 03:00:00 266.0 114.0 \n", + "2015-01-01 04:00:00 204.0 122.0 \n", + "2015-01-01 05:00:00 66.0 72.0 \n", + "2015-01-01 06:00:00 26.0 70.0 \n", + "2015-01-01 07:00:00 13.0 57.0 \n", + "2015-01-01 08:00:00 10.0 31.0 \n", + "2015-01-01 09:00:00 12.0 27.0 \n", + "2015-01-01 10:00:00 13.0 16.0 \n", + "2015-01-01 11:00:00 10.0 11.0 \n", + "2015-01-01 12:00:00 11.0 24.0 \n", + "2015-01-01 13:00:00 23.0 13.0 \n", + "2015-01-01 14:00:00 31.0 18.0 \n", + "2015-01-01 15:00:00 32.0 24.0 \n", + "2015-01-01 16:00:00 43.0 24.0 \n", + "2015-01-01 17:00:00 37.0 25.0 \n", + "2015-01-01 18:00:00 38.0 26.0 \n", + "2015-01-01 19:00:00 44.0 21.0 \n", + "2015-01-01 20:00:00 30.0 27.0 \n", + "2015-01-01 21:00:00 43.0 23.0 \n", + "2015-01-01 22:00:00 38.0 29.0 \n", + "2015-01-01 23:00:00 39.0 27.0 \n", + "2015-01-02 00:00:00 27.0 19.0 \n", + "2015-01-02 01:00:00 31.0 16.0 \n", + "2015-01-02 02:00:00 15.0 9.0 \n", + "2015-01-02 03:00:00 30.0 9.0 \n", + "2015-01-02 04:00:00 28.0 12.0 \n", + "2015-01-02 05:00:00 13.0 6.0 \n", + "2015-01-02 06:00:00 9.0 4.0 \n", + "... ... ... \n", + "2015-06-29 18:00:00 43.0 36.0 \n", + "2015-06-29 19:00:00 52.0 35.0 \n", + "2015-06-29 20:00:00 50.0 41.0 \n", + "2015-06-29 21:00:00 54.0 45.0 \n", + "2015-06-29 22:00:00 49.0 49.0 \n", + "2015-06-29 23:00:00 52.0 50.0 \n", + "2015-06-30 00:00:00 44.0 34.0 \n", + "2015-06-30 01:00:00 34.0 17.0 \n", + "2015-06-30 02:00:00 12.0 12.0 \n", + "2015-06-30 03:00:00 9.0 9.0 \n", + "2015-06-30 04:00:00 11.0 11.0 \n", + "2015-06-30 05:00:00 15.0 10.0 \n", + "2015-06-30 06:00:00 26.0 15.0 \n", + "2015-06-30 07:00:00 38.0 32.0 \n", + "2015-06-30 08:00:00 50.0 37.0 \n", + "2015-06-30 09:00:00 58.0 35.0 \n", + "2015-06-30 10:00:00 44.0 37.0 \n", + "2015-06-30 11:00:00 36.0 32.0 \n", + "2015-06-30 12:00:00 33.0 25.0 \n", + "2015-06-30 13:00:00 30.0 13.0 \n", + "2015-06-30 14:00:00 28.0 25.0 \n", + "2015-06-30 15:00:00 34.0 20.0 \n", + "2015-06-30 16:00:00 34.0 25.0 \n", + "2015-06-30 17:00:00 53.0 26.0 \n", + "2015-06-30 18:00:00 43.0 41.0 \n", + "2015-06-30 19:00:00 72.0 49.0 \n", + "2015-06-30 20:00:00 64.0 61.0 \n", + "2015-06-30 21:00:00 77.0 50.0 \n", + "2015-06-30 22:00:00 89.0 58.0 \n", + "2015-06-30 23:00:00 105.0 61.0 \n", + "\n", + " Windsor Terrace Woodhaven Woodside World Trade Center \\\n", + "date \n", + "2015-01-01 01:00:00 15.0 5.0 2.0 18.0 \n", + "2015-01-01 02:00:00 7.0 9.0 4.0 8.0 \n", + "2015-01-01 03:00:00 9.0 12.0 10.0 11.0 \n", + "2015-01-01 04:00:00 8.0 7.0 11.0 15.0 \n", + "2015-01-01 05:00:00 2.0 6.0 6.0 5.0 \n", + "2015-01-01 06:00:00 2.0 1.0 4.0 6.0 \n", + "2015-01-01 07:00:00 2.0 1.0 3.0 3.0 \n", + "2015-01-01 08:00:00 3.0 2.0 0.0 3.0 \n", + "2015-01-01 09:00:00 2.0 1.0 5.0 2.0 \n", + "2015-01-01 10:00:00 0.0 3.0 1.0 3.0 \n", + "2015-01-01 11:00:00 3.0 3.0 1.0 1.0 \n", + "2015-01-01 12:00:00 6.0 1.0 1.0 8.0 \n", + "2015-01-01 13:00:00 2.0 1.0 7.0 6.0 \n", + "2015-01-01 14:00:00 1.0 1.0 8.0 8.0 \n", + "2015-01-01 15:00:00 2.0 4.0 1.0 12.0 \n", + "2015-01-01 16:00:00 9.0 3.0 0.0 17.0 \n", + "2015-01-01 17:00:00 7.0 3.0 5.0 18.0 \n", + "2015-01-01 18:00:00 7.0 2.0 2.0 22.0 \n", + "2015-01-01 19:00:00 3.0 0.0 3.0 18.0 \n", + "2015-01-01 20:00:00 3.0 0.0 2.0 8.0 \n", + "2015-01-01 21:00:00 5.0 0.0 2.0 8.0 \n", + "2015-01-01 22:00:00 2.0 2.0 0.0 10.0 \n", + "2015-01-01 23:00:00 1.0 1.0 4.0 3.0 \n", + "2015-01-02 00:00:00 0.0 3.0 2.0 0.0 \n", + "2015-01-02 01:00:00 0.0 1.0 2.0 2.0 \n", + "2015-01-02 02:00:00 2.0 1.0 1.0 3.0 \n", + "2015-01-02 03:00:00 0.0 0.0 0.0 2.0 \n", + "2015-01-02 04:00:00 2.0 1.0 1.0 1.0 \n", + "2015-01-02 05:00:00 1.0 0.0 1.0 1.0 \n", + "2015-01-02 06:00:00 4.0 1.0 0.0 1.0 \n", + "... ... ... ... ... \n", + "2015-06-29 18:00:00 3.0 0.0 7.0 30.0 \n", + "2015-06-29 19:00:00 10.0 2.0 2.0 41.0 \n", + "2015-06-29 20:00:00 3.0 5.0 4.0 34.0 \n", + "2015-06-29 21:00:00 5.0 4.0 6.0 28.0 \n", + "2015-06-29 22:00:00 5.0 2.0 10.0 17.0 \n", + "2015-06-29 23:00:00 2.0 3.0 6.0 25.0 \n", + "2015-06-30 00:00:00 4.0 1.0 5.0 15.0 \n", + "2015-06-30 01:00:00 3.0 0.0 0.0 10.0 \n", + "2015-06-30 02:00:00 2.0 0.0 1.0 7.0 \n", + "2015-06-30 03:00:00 0.0 1.0 4.0 2.0 \n", + "2015-06-30 04:00:00 0.0 0.0 2.0 8.0 \n", + "2015-06-30 05:00:00 1.0 3.0 1.0 7.0 \n", + "2015-06-30 06:00:00 4.0 4.0 5.0 10.0 \n", + "2015-06-30 07:00:00 6.0 3.0 6.0 13.0 \n", + "2015-06-30 08:00:00 10.0 6.0 8.0 20.0 \n", + "2015-06-30 09:00:00 7.0 5.0 6.0 16.0 \n", + "2015-06-30 10:00:00 5.0 2.0 7.0 14.0 \n", + "2015-06-30 11:00:00 5.0 1.0 2.0 21.0 \n", + "2015-06-30 12:00:00 4.0 5.0 3.0 15.0 \n", + "2015-06-30 13:00:00 5.0 3.0 5.0 16.0 \n", + "2015-06-30 14:00:00 7.0 2.0 8.0 24.0 \n", + "2015-06-30 15:00:00 4.0 2.0 4.0 16.0 \n", + "2015-06-30 16:00:00 4.0 4.0 7.0 32.0 \n", + "2015-06-30 17:00:00 4.0 3.0 4.0 39.0 \n", + "2015-06-30 18:00:00 5.0 8.0 3.0 39.0 \n", + "2015-06-30 19:00:00 8.0 4.0 10.0 43.0 \n", + "2015-06-30 20:00:00 10.0 4.0 6.0 42.0 \n", + "2015-06-30 21:00:00 4.0 2.0 6.0 24.0 \n", + "2015-06-30 22:00:00 6.0 2.0 9.0 34.0 \n", + "2015-06-30 23:00:00 5.0 1.0 12.0 16.0 \n", + "\n", + " Yorkville East Yorkville West temp precip \n", + "date \n", + "2015-01-01 01:00:00 39.0 83.0 0.298851 0.0 \n", + "2015-01-01 02:00:00 27.0 51.0 0.298851 0.0 \n", + "2015-01-01 03:00:00 20.0 42.0 0.287356 0.0 \n", + "2015-01-01 04:00:00 19.0 28.0 0.287356 0.0 \n", + "2015-01-01 05:00:00 7.0 13.0 0.287356 0.0 \n", + "2015-01-01 06:00:00 6.0 9.0 0.287356 0.0 \n", + "2015-01-01 07:00:00 3.0 8.0 0.287356 0.0 \n", + "2015-01-01 08:00:00 3.0 2.0 0.287356 0.0 \n", + "2015-01-01 09:00:00 5.0 7.0 0.298851 0.0 \n", + "2015-01-01 10:00:00 8.0 5.0 0.321839 0.0 \n", + "2015-01-01 11:00:00 11.0 14.0 0.333333 0.0 \n", + "2015-01-01 12:00:00 11.0 14.0 0.356322 0.0 \n", + "2015-01-01 13:00:00 8.0 17.0 0.379310 0.0 \n", + "2015-01-01 14:00:00 9.0 19.0 0.402299 0.0 \n", + "2015-01-01 15:00:00 23.0 22.0 0.413793 0.0 \n", + "2015-01-01 16:00:00 12.0 24.0 0.425287 0.0 \n", + "2015-01-01 17:00:00 14.0 30.0 0.425287 0.0 \n", + "2015-01-01 18:00:00 13.0 37.0 0.425287 0.0 \n", + "2015-01-01 19:00:00 13.0 19.0 0.425287 0.0 \n", + "2015-01-01 20:00:00 22.0 26.0 0.413793 0.0 \n", + "2015-01-01 21:00:00 19.0 19.0 0.413793 0.0 \n", + "2015-01-01 22:00:00 11.0 20.0 0.413793 0.0 \n", + "2015-01-01 23:00:00 9.0 24.0 0.413793 0.0 \n", + "2015-01-02 00:00:00 4.0 13.0 0.402299 0.0 \n", + "2015-01-02 01:00:00 4.0 14.0 0.413793 0.0 \n", + "2015-01-02 02:00:00 3.0 6.0 0.402299 0.0 \n", + "2015-01-02 03:00:00 2.0 5.0 0.402299 0.0 \n", + "2015-01-02 04:00:00 1.0 6.0 0.402299 0.0 \n", + "2015-01-02 05:00:00 9.0 7.0 0.390805 0.0 \n", + "2015-01-02 06:00:00 5.0 9.0 0.390805 0.0 \n", + "... ... ... ... ... \n", + "2015-06-29 18:00:00 36.0 19.0 0.839080 0.0 \n", + "2015-06-29 19:00:00 25.0 30.0 0.839080 0.0 \n", + "2015-06-29 20:00:00 20.0 28.0 0.827586 0.0 \n", + "2015-06-29 21:00:00 18.0 27.0 0.816092 0.0 \n", + "2015-06-29 22:00:00 13.0 35.0 0.804598 0.0 \n", + "2015-06-29 23:00:00 18.0 16.0 0.804598 0.0 \n", + "2015-06-30 00:00:00 7.0 16.0 0.804598 0.0 \n", + "2015-06-30 01:00:00 5.0 7.0 0.804598 0.0 \n", + "2015-06-30 02:00:00 4.0 3.0 0.793103 0.0 \n", + "2015-06-30 03:00:00 3.0 1.0 0.781609 0.0 \n", + "2015-06-30 04:00:00 4.0 8.0 0.770115 0.0 \n", + "2015-06-30 05:00:00 9.0 13.0 0.758621 0.0 \n", + "2015-06-30 06:00:00 18.0 18.0 0.781609 0.0 \n", + "2015-06-30 07:00:00 34.0 40.0 0.816092 0.0 \n", + "2015-06-30 08:00:00 68.0 63.0 0.839080 0.0 \n", + "2015-06-30 09:00:00 84.0 71.0 0.862069 0.0 \n", + "2015-06-30 10:00:00 64.0 51.0 0.873563 0.0 \n", + "2015-06-30 11:00:00 38.0 31.0 0.885057 0.0 \n", + "2015-06-30 12:00:00 28.0 28.0 0.885057 0.0 \n", + "2015-06-30 13:00:00 22.0 21.0 0.908046 0.0 \n", + "2015-06-30 14:00:00 14.0 26.0 0.896552 0.0 \n", + "2015-06-30 15:00:00 13.0 28.0 0.896552 0.0 \n", + "2015-06-30 16:00:00 23.0 28.0 0.850575 0.0 \n", + "2015-06-30 17:00:00 25.0 30.0 0.850575 0.0 \n", + "2015-06-30 18:00:00 42.0 36.0 0.839080 0.0 \n", + "2015-06-30 19:00:00 39.0 41.0 0.835249 0.0 \n", + "2015-06-30 20:00:00 34.0 34.0 0.823755 0.0 \n", + "2015-06-30 21:00:00 24.0 39.0 0.827586 0.0 \n", + "2015-06-30 22:00:00 25.0 38.0 0.827586 0.0 \n", + "2015-06-30 23:00:00 14.0 34.0 0.827586 0.0 \n", + "\n", + "[4343 rows x 142 columns]" + ] + }, + "execution_count": 3, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "rides_weather.head()" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + }, + { + "cell_type": "code", + "execution_count": 4, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "tensorflow version 1.8.0\n", + "pandas version 0.22.0\n" + ] + } + ], + "source": [ + "print(\"tensorflow version\", tf.__version__)\n", + "#print(\"keras version\", tf.keras.__version__)\n", + "print(\"pandas version\", pd.__version__)" + ] + }, + { + "cell_type": "code", + "execution_count": 311, + "metadata": {}, + "outputs": [ + { + "ename": "AttributeError", + "evalue": "'DatetimeIndex' object has no attribute 'weekend'", + "output_type": "error", + "traceback": [ + "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", + "\u001b[0;31mAttributeError\u001b[0m Traceback (most recent call last)", + "\u001b[0;32m\u001b[0m in \u001b[0;36m\u001b[0;34m()\u001b[0m\n\u001b[1;32m 2\u001b[0m \u001b[0;31m# df['Various', 'Day'] = df.index.dayofyear\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 3\u001b[0m \u001b[0;31m# df['Various', 'Hour'] = df.index.hour\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0mrides_weather\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mweekend\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m", + "\u001b[0;31mAttributeError\u001b[0m: 'DatetimeIndex' object has no attribute 'weekend'" + ] + } + ], + "source": [ + "# potentially add day and hour later\n", + "# df['Various', 'Day'] = df.index.dayofyear\n", + "# df['Various', 'Hour'] = df.index.hour\n", + "# rides_weather.index.weekend" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": {}, + "outputs": [], + "source": [ + "target_names = rides_weather.columns[:-2]" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "We are going to predict 24hours into the future" + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": {}, + "outputs": [], + "source": [ + "shift_days = 1\n", + "shift_steps = shift_days * 24 # Number of hours to shift." + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "metadata": {}, + "outputs": [], + "source": [ + "# shifts the time so that we predict 24 hours into the future\n", + "df_targets = rides_weather[target_names].shift(-shift_steps)" + ] + }, + { + "cell_type": "code", + "execution_count": 60, + "metadata": {}, + "outputs": [], + "source": [ + "x_data = rides_weather[target_names].values[0:-shift_steps]\n", + "y_data = df_targets.values[:-shift_steps]\n", + "num_data = len(x_data)" + ] + }, + { + "cell_type": "code", + "execution_count": 61, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "X_data Shape: (4319, 140)\n", + "y_data Shape: (4319, 140)\n" + ] + } + ], + "source": [ + "print(\"X_data Shape:\", x_data.shape)\n", + "print(\"y_data Shape:\", y_data.shape)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "We will split the data so that our test set will be the last week of June. The data will start being collected from the week prior to the test week." + ] + }, + { + "cell_type": "code", + "execution_count": 62, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "length x_train: 3983, y_train: 3983\n", + "length x_test: 336, y_test: 336\n" + ] + } + ], + "source": [ + "# Split test data, which will be last week of June, but will be trained on the 2nd to last week of June\n", + "# 24hrs * 2 weeks\n", + "num_test = 24*7*2\n", + "num_train_set = num_data - num_test\n", + "x_train_set = x_data[0:num_train_set]\n", + "x_test = x_data[num_train_set:]\n", + "y_train_set = y_data[0:num_train_set]\n", + "y_test = y_data[num_train_set:]\n", + "print(\"length x_train: {}, y_train: {}\".format(len(x_train_set),len(y_train_set)))\n", + "print(\"length x_test: {}, y_test: {}\".format(len(x_test),len(y_test)))" + ] + }, + { + "cell_type": "code", + "execution_count": 63, + "metadata": {}, + "outputs": [], + "source": [ + "# We split the training set further into training and cross validation set\n", + "x_train, x_val, y_train, y_val = train_test_split(x_train_set, y_train_set, test_size=0.15, shuffle=False)" + ] + }, + { + "cell_type": "code", + "execution_count": 93, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "length x_train: 3385, y_train: 3385\n", + "length x_test: 598, y_test: 598\n" + ] + } + ], + "source": [ + "num_train = len(x_train)\n", + "print(\"length x_train: {}, y_train: {}\".format(len(x_train),len(y_train)))\n", + "print(\"length x_test: {}, y_test: {}\".format(len(x_val),len(y_val)))" + ] + }, + { + "cell_type": "code", + "execution_count": 69, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "num x features: 140, num y features: 140\n" + ] + } + ], + "source": [ + "num_x_signals = x_data.shape[1]\n", + "num_y_signals = y_data.shape[1]\n", + "print(\"num x features: {}, num y features: {}\".format(num_x_signals,num_y_signals))" + ] + }, + { + "cell_type": "code", + "execution_count": 75, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "for x_train: Min: 0.0, Max: 597.0\n", + "for x_val: Min: 0.0, Max: 576.0\n", + "for y_test: Min: 0.0, Max: 569.0\n", + "The max output for test set will be whatever the max value is for x_train\n" + ] + } + ], + "source": [ + "print(\"for x_train: Min: {}, Max: {}\".format(np.min(x_train),np.max(x_train)))\n", + "print(\"for x_val: Min: {}, Max: {}\".format(np.min(x_val),np.max(x_val)))\n", + "print(\"for y_test: Min: {}, Max: {}\".format(np.min(x_test),np.max(x_test)))\n", + "print(\"The max output for test set will be whatever the max value is for x_train\")" + ] + }, + { + "cell_type": "code", + "execution_count": 117, + "metadata": {}, + "outputs": [], + "source": [ + "# Apply mim max scaler to all values from 0 to 1\n", + "x_scaler = MinMaxScaler()\n", + "x_train_scaled = x_scaler.fit_transform(x_train)\n", + "x_val_scaled = x_scaler.transform(x_val)\n", + "x_test_scaled = x_scaler.transform(x_test)\n", + "y_scaler = MinMaxScaler()\n", + "y_train_scaled = y_scaler.fit_transform(y_train)\n", + "y_val_scaled = y_scaler.transform(y_val)\n", + "y_test_scaled = y_scaler.transform(y_test)" + ] + }, + { + "cell_type": "code", + "execution_count": 77, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "(3385, 140)\n", + "(3385, 140)\n" + ] + } + ], + "source": [ + "print(x_train_scaled.shape)\n", + "print(y_train_scaled.shape)" + ] + }, + { + "cell_type": "code", + "execution_count": 97, + "metadata": {}, + "outputs": [], + "source": [ + "def batch_generator(batch_size, sequence_length):\n", + " \"\"\"\n", + " Generator function for creating random batches of training-data.\n", + " \"\"\"\n", + "\n", + " while True:\n", + " # Allocate a new array for the batch of input-signals.\n", + " x_shape = (batch_size, sequence_length, num_x_signals)\n", + " x_batch = np.zeros(shape=x_shape, dtype=np.float16)\n", + "\n", + " # Allocate a new array for the batch of output-signals.\n", + " y_shape = (batch_size, sequence_length, num_y_signals)\n", + " y_batch = np.zeros(shape=y_shape, dtype=np.float16)\n", + "\n", + " # Fill the batch with random sequences of data.\n", + " for i in range(batch_size):\n", + " # Get a random start-index.\n", + " # This points somewhere into the training-data.\n", + " idx = np.random.randint(num_train - sequence_length)\n", + " \n", + " # Copy the sequences of data starting at this index.\n", + " x_batch[i] = x_train_scaled[idx:idx+sequence_length]\n", + " y_batch[i] = y_train_scaled[idx:idx+sequence_length]\n", + " \n", + " yield (x_batch, y_batch)" + ] + }, + { + "cell_type": "code", + "execution_count": 98, + "metadata": {}, + "outputs": [], + "source": [ + "# Sequence length will be a week's worth of data\n", + "batch_size = 256\n", + "sequence_length = 24 * 7\n", + "generator = batch_generator(batch_size=batch_size,\n", + " sequence_length=sequence_length)" + ] + }, + { + "cell_type": "code", + "execution_count": 99, + "metadata": {}, + "outputs": [], + "source": [ + "x_batch, y_batch = next(generator)" + ] + }, + { + "cell_type": "code", + "execution_count": 100, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "(256, 168, 140)\n", + "(256, 168, 140)\n" + ] + } + ], + "source": [ + "print(x_batch.shape)\n", + "print(y_batch.shape)" + ] + }, + { + "cell_type": "code", + "execution_count": 101, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "[]" + ] + }, + "execution_count": 101, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "batch = 0 # First sequence in the batch.\n", + "signal = 0 # First signal from the 20 input-signals.\n", + "seq = x_batch[batch, :, signal]\n", + "plt.plot(seq)" + ] + }, + { + "cell_type": "code", + "execution_count": 102, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "[]" + ] + }, + "execution_count": 102, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "seq = y_batch[batch, :, signal]\n", + "plt.plot(seq)" + ] + }, + { + "cell_type": "code", + "execution_count": 103, + "metadata": {}, + "outputs": [], + "source": [ + "validation_data = (np.expand_dims(x_val_scaled, axis=0),\n", + " np.expand_dims(y_val_scaled, axis=0))" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Build the model" + ] + }, + { + "cell_type": "code", + "execution_count": 106, + "metadata": {}, + "outputs": [], + "source": [ + "# Define loss function\n", + "\n", + "# warmup steps is the timeframe which won't be penalized/considered, \n", + "# allowing the model to have sufficient information to make its prediction\n", + "warmup_steps=24\n", + "\n", + "def loss_mse_warmup(y_true, y_pred):\n", + " \"\"\"\n", + " Calculate the Mean Squared Error between y_true and y_pred,\n", + " but ignore the beginning \"warmup\" part of the sequences, where\n", + " the prediction will be poor.\n", + " \n", + " y_true is the desired output.\n", + " y_pred is the model's output.\n", + " \"\"\"\n", + "\n", + " # The shape of both input tensors are:\n", + " # [batch_size, sequence_length, num_y_signals].\n", + "\n", + " # Ignore the \"warmup\" parts of the sequences\n", + " # by taking slices of the tensors.\n", + " y_true_slice = y_true[:, warmup_steps:, :]\n", + " y_pred_slice = y_pred[:, warmup_steps:, :]\n", + "\n", + " # These sliced tensors both have this shape:\n", + " # [batch_size, sequence_length - warmup_steps, num_y_signals]\n", + "\n", + " # Calculate the MSE loss for each value in these tensors.\n", + " # This outputs a 3-rank tensor of the same shape.\n", + " loss = tf.losses.mean_squared_error(labels=y_true_slice,\n", + " predictions=y_pred_slice)\n", + "\n", + " # Keras may reduce this across the first axis (the batch)\n", + " # but the semantics are unclear, so to be sure we use\n", + " # the loss across the entire tensor, we reduce it to a\n", + " # single scalar with the mean function.\n", + " loss_mean = tf.reduce_mean(loss)\n", + "\n", + " return loss_mean" + ] + }, + { + "cell_type": "code", + "execution_count": 294, + "metadata": {}, + "outputs": [], + "source": [ + "def build_model_RCU(batch_size = 256, sequence_length = 24*7):\n", + " '''\n", + " Builds a shallow 1 layer GRU network.\n", + " Sequence length can be specified if other than 1 week\n", + " '''\n", + " \n", + " # generate batches with specified batch size & sequence length\n", + " # default is at 1 week\n", + " generator = batch_generator(batch_size=batch_size,\n", + " sequence_length=sequence_length)\n", + " \n", + " #Model architecture\n", + " model = Sequential()\n", + "\n", + " # Add a GRU layer of 256 length (matching batch size)\n", + " model.add(GRU(units=256,return_sequences=True,input_shape=(None,num_x_signals)))\n", + "\n", + " # Add a dense layer for output\n", + " model.add(Dense(num_y_signals, activation='sigmoid'))\n", + " \n", + " optimizer = RMSprop(lr=1e-3)\n", + " model.compile(loss=loss_mse_warmup, optimizer=optimizer)\n", + " print(model.summary())\n", + " \n", + " # Define challbacks\n", + " path_checkpoint = '23_checkpoint.keras'\n", + " callback_checkpoint = ModelCheckpoint(filepath=path_checkpoint,\n", + " monitor='val_loss',\n", + " verbose=1,\n", + " save_weights_only=True,\n", + " save_best_only=True)\n", + " \n", + " # stop the model if the cross validation has not improved after\n", + " # 5 successive epochs\n", + " callback_early_stopping = EarlyStopping(monitor='val_loss',\n", + " patience=5, verbose=1)\n", + " \n", + " # log the \n", + " callback_tensorboard = TensorBoard(log_dir='./23_logs/',\n", + " histogram_freq=0,\n", + " write_graph=False)\n", + " \n", + " # if cross validation score decreases from previous epoch, then\n", + " # immediately lower the learning rate by a factor of 10\n", + " callback_reduce_lr = ReduceLROnPlateau(monitor='val_loss',\n", + " factor=0.1,\n", + " min_lr=1e-4,\n", + " patience=0,\n", + " verbose=1)\n", + " callbacks = [callback_early_stopping,\n", + " callback_checkpoint,\n", + " callback_tensorboard,\n", + " callback_reduce_lr]\n", + " return generator, model, callbacks" + ] + }, + { + "cell_type": "code", + "execution_count": 297, + "metadata": {}, + "outputs": [], + "source": [ + "def fit_model(batch_size = 256, sequence_length = 24*7):\n", + " generator, model, callbacks = build_model_RCU(batch_size, sequence_length)\n", + " %%time\n", + " model.fit_generator(generator=generator,\n", + " epochs=20,\n", + " steps_per_epoch=100,\n", + " validation_data=validation_data,\n", + " callbacks=callbacks)\n", + " return model" + ] + }, + { + "cell_type": "code", + "execution_count": 298, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "_________________________________________________________________\n", + "Layer (type) Output Shape Param # \n", + "=================================================================\n", + "gru_3 (GRU) (None, None, 256) 304896 \n", + "_________________________________________________________________\n", + "dense_3 (Dense) (None, None, 140) 35980 \n", + "=================================================================\n", + "Total params: 340,876\n", + "Trainable params: 340,876\n", + "Non-trainable params: 0\n", + "_________________________________________________________________\n", + "None\n", + "CPU times: user 0 ns, sys: 0 ns, total: 0 ns\n", + "Wall time: 7.39 µs\n", + "Epoch 1/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0205\n", + "Epoch 00001: val_loss improved from inf to 0.02359, saving model to 23_checkpoint.keras\n", + "100/100 [==============================] - 71s 706ms/step - loss: 0.0205 - val_loss: 0.0236\n", + "Epoch 2/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0100\n", + "Epoch 00002: val_loss improved from 0.02359 to 0.01669, saving model to 23_checkpoint.keras\n", + "100/100 [==============================] - 69s 693ms/step - loss: 0.0100 - val_loss: 0.0167\n", + "Epoch 3/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0087\n", + "Epoch 00003: val_loss improved from 0.01669 to 0.01603, saving model to 23_checkpoint.keras\n", + "100/100 [==============================] - 69s 691ms/step - loss: 0.0086 - val_loss: 0.0160\n", + "Epoch 4/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0079\n", + "Epoch 00004: val_loss did not improve\n", + "\n", + "Epoch 00004: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n", + "100/100 [==============================] - 69s 690ms/step - loss: 0.0079 - val_loss: 0.0163\n", + "Epoch 5/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0069\n", + "Epoch 00005: val_loss improved from 0.01603 to 0.01455, saving model to 23_checkpoint.keras\n", + "100/100 [==============================] - 69s 695ms/step - loss: 0.0069 - val_loss: 0.0146\n", + "Epoch 6/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0067\n", + "Epoch 00006: val_loss improved from 0.01455 to 0.01430, saving model to 23_checkpoint.keras\n", + "100/100 [==============================] - 70s 696ms/step - loss: 0.0067 - val_loss: 0.0143\n", + "Epoch 7/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0065\n", + "Epoch 00007: val_loss did not improve\n", + "\n", + "Epoch 00007: ReduceLROnPlateau reducing learning rate to 0.0001.\n", + "100/100 [==============================] - 69s 689ms/step - loss: 0.0065 - val_loss: 0.0144\n", + "Epoch 8/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0064\n", + "Epoch 00008: val_loss did not improve\n", + "100/100 [==============================] - 69s 688ms/step - loss: 0.0064 - val_loss: 0.0144\n", + "Epoch 9/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0063\n", + "Epoch 00009: val_loss did not improve\n", + "100/100 [==============================] - 69s 695ms/step - loss: 0.0063 - val_loss: 0.0144\n", + "Epoch 10/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0063\n", + "Epoch 00010: val_loss did not improve\n", + "100/100 [==============================] - 70s 696ms/step - loss: 0.0063 - val_loss: 0.0144\n", + "Epoch 11/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0062\n", + "Epoch 00011: val_loss did not improve\n", + "100/100 [==============================] - 69s 694ms/step - loss: 0.0062 - val_loss: 0.0145\n", + "Epoch 00011: early stopping\n" + ] + } + ], + "source": [ + "model2 = fit_model(sequence_length = 24*14)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Train the model" + ] + }, + { + "cell_type": "code", + "execution_count": 113, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Epoch 1/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0178\n", + "Epoch 00001: val_loss improved from inf to 0.02102, saving model to 23_checkpoint.keras\n", + "100/100 [==============================] - 35s 347ms/step - loss: 0.0177 - val_loss: 0.0210\n", + "Epoch 2/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0099\n", + "Epoch 00002: val_loss improved from 0.02102 to 0.01714, saving model to 23_checkpoint.keras\n", + "100/100 [==============================] - 34s 338ms/step - loss: 0.0099 - val_loss: 0.0171\n", + "Epoch 3/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0086\n", + "Epoch 00003: val_loss improved from 0.01714 to 0.01433, saving model to 23_checkpoint.keras\n", + "100/100 [==============================] - 34s 340ms/step - loss: 0.0086 - val_loss: 0.0143\n", + "Epoch 4/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0080\n", + "Epoch 00004: val_loss did not improve\n", + "\n", + "Epoch 00004: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n", + "100/100 [==============================] - 34s 343ms/step - loss: 0.0080 - val_loss: 0.0171\n", + "Epoch 5/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0071\n", + "Epoch 00005: val_loss did not improve\n", + "\n", + "Epoch 00005: ReduceLROnPlateau reducing learning rate to 0.0001.\n", + "100/100 [==============================] - 34s 337ms/step - loss: 0.0071 - val_loss: 0.0145\n", + "Epoch 6/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0068\n", + "Epoch 00006: val_loss did not improve\n", + "100/100 [==============================] - 34s 339ms/step - loss: 0.0068 - val_loss: 0.0144\n", + "Epoch 7/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0067\n", + "Epoch 00007: val_loss did not improve\n", + "100/100 [==============================] - 34s 340ms/step - loss: 0.0067 - val_loss: 0.0144\n", + "Epoch 8/20\n", + " 99/100 [============================>.] - ETA: 0s - loss: 0.0066\n", + "Epoch 00008: val_loss did not improve\n", + "100/100 [==============================] - 34s 339ms/step - loss: 0.0066 - val_loss: 0.0146\n", + "Epoch 00008: early stopping\n", + "CPU times: user 6min 47s, sys: 47.8 s, total: 7min 35s\n", + "Wall time: 4min 33s\n" + ] + }, + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 113, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "%%time\n", + "model.fit_generator(generator=generator,\n", + " epochs=20,\n", + " steps_per_epoch=100,\n", + " validation_data=validation_data,\n", + " callbacks=callbacks)" + ] + }, + { + "cell_type": "code", + "execution_count": 300, + "metadata": {}, + "outputs": [], + "source": [ + "def predict_future(model_name, x_scaled):\n", + " '''\n", + " Returns the prediction given scaled x test\n", + " '''\n", + " x = np.expand_dims(x_scaled, axis=0)\n", + " y_pred = model_name.predict(x)\n", + " y_pred_rescaled = y_scaler.inverse_transform(y_pred[0])\n", + " return y_pred_rescaled" + ] + }, + { + "cell_type": "code", + "execution_count": 253, + "metadata": {}, + "outputs": [], + "source": [ + "# Define metric for comparing models\n", + "def calc_rmse_total(prediction_length, y_true, y_pred):\n", + " \"\"\"\n", + " Calculate the Mean Squared Error between y_true and y_pred\n", + " across all the locations, then take the square root to get rmse.\n", + " Ignore the training week.\n", + " \n", + " prediction length: the length of time that you want to predict\n", + " y_true[time, location] is the desired output.\n", + " y_pred[time, location] is the model's output.\n", + " \"\"\"\n", + "\n", + " # Ignore the \"the training week\"\n", + " y_true_slice = y_true[-prediction_length:, :]\n", + " y_pred_slice = y_pred[-prediction_length:, :]\n", + "\n", + " # Calculate the MSE loss for each value in these tensors.\n", + " # This outputs a 3-rank tensor of the same shape.\n", + " mse_list = []\n", + " for i in range(len(y_true_slice.T)):\n", + " mse_list.append(mean_squared_error(y_true_slice.T[i], y_pred_slice.T[i]))\n", + " rmse_total = sqrt(sum(mse_list))\n", + "\n", + " return rmse_total" + ] + }, + { + "cell_type": "code", + "execution_count": 169, + "metadata": {}, + "outputs": [], + "source": [ + "# Save model\n", + "# model.save(\"model1.model\")" + ] + }, + { + "cell_type": "code", + "execution_count": 115, + "metadata": {}, + "outputs": [], + "source": [ + "try:\n", + " model.load_weights(path_checkpoint)\n", + "except Exception as error:\n", + " print(\"Error trying to load checkpoint.\")\n", + " print(error)" + ] + }, + { + "cell_type": "code", + "execution_count": 119, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\r", + "1/1 [==============================] - 0s 105ms/step\n" + ] + } + ], + "source": [ + "result = model.evaluate(x=np.expand_dims(x_test_scaled, axis=0),\n", + " y=np.expand_dims(y_test_scaled, axis=0))\n", + "print(\"loss (test-set):\", result)" + ] + }, + { + "cell_type": "code", + "execution_count": 306, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "model RMSE: 185.73497789219232\n", + "model2 RMSE: 171.56668638399555\n", + "RMSE improves when we have a longer sequence length for batch training (2weeks)\n" + ] + } + ], + "source": [ + "# calculate rmse across all locations for the last week of June\n", + "y_pred_rescaled = predict_future(model,x_test_scaled)\n", + "y_pred_rescaled2 = predict_future(model2,x_test_scaled)\n", + "print(\"model (1week sequence) RMSE:\", calc_rmse_total(24*7, y_test, y_pred_rescaled))\n", + "print(\"model2 (2weeks sequence) RMSE:\", calc_rmse_total(24*7, y_test, y_pred_rescaled2))\n", + "print(\"RMSE improves when we have a longer sequence length for batch training (2weeks)\")" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### total RMSE for the last week of June came out to be 185.7 rides" + ] + }, + { + "cell_type": "code", + "execution_count": 256, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 256, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "plt.figure(figsize=(15,5))\n", + "plt.plot(y_pred_rescaled[:,1],label=\"pred\");\n", + "plt.plot(y_test[:,1],label=\"actual\");\n", + "plt.legend()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Let's make sure that the prediction is the same with varying sequence length as long as my starting index is kept the same" + ] + }, + { + "cell_type": "code", + "execution_count": 292, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "The shaded area indicates training phase. White area is the prediction phase\n" + ] + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "y_pred_rescaled1 = predict_future(x_test_scaled[:24*8])\n", + "y_pred_rescaled2 = predict_future(x_test_scaled[:24*9])\n", + "y_pred_rescaled3 = predict_future(x_test_scaled[:24*10])\n", + "\n", + "print(\"The shaded area indicates training phase. White area is the prediction phase.\")\n", + "print(\"The plots are to demonstrate that if I were to predict \")\n", + "\n", + "f = plt.figure(figsize=(10,10));\n", + "ax1 = f.add_subplot(311)\n", + "ax2 = f.add_subplot(312)\n", + "ax3 = f.add_subplot(313)\n", + "ax1.plot(y_pred_rescaled1[:,1],label=\"pred\");\n", + "#ax1.plot(x_test[:24*8,1],label=\"x\");\n", + "ax2.plot(y_pred_rescaled2[:,1],label=\"pred2\");\n", + "ax3.plot(y_pred_rescaled3[:,1],label=\"pred3\");\n", + "ax3.plot(y_test[:,1],label=\"actual\");\n", + "ax1.set_title(\"Predict 1 day\")\n", + "ax2.set_title(\"Predict 2 days\")\n", + "ax3.set_title(\"Predict 3 days\")\n", + "ax1.axvspan(0,24*7,facecolor=\"black\",alpha=0.15)\n", + "ax2.axvspan(0,24*7,facecolor=\"black\",alpha=0.15)\n", + "ax3.axvspan(0,24*7,facecolor=\"black\",alpha=0.15)\n", + "ax1.set_xlim([0, 250])\n", + "ax2.set_xlim([0, 250])\n", + "ax3.set_xlim([0, 250])\n", + "\n", + "plt.legend();" + ] + }, + { + "cell_type": "code", + "execution_count": 281, + "metadata": {}, + "outputs": [], + "source": [ + "import random\n", + "def plot_last_week(prediction_length=24*7, train=False):\n", + " \"\"\"\n", + " Plot the prediction for last week of June.\n", + "\n", + " prediction_length = last week of june (24*7hrs)\n", + " train: if true, will look at last week of training data\n", + " \"\"\"\n", + " \n", + " if train:\n", + " # Use training-data.\n", + " x = x_train_scaled\n", + " y_true = y_train\n", + " else:\n", + " # Use test-data.\n", + " x = x_test_scaled\n", + " y_true = y_test\n", + " \n", + " y_pred_rescaled = predict_future(x)\n", + " \n", + " # plot 3 random locations\n", + " for i in range(3):\n", + " # Get the output-signal predicted by the model.\n", + " signal = random.randint(0,139)\n", + " signal_pred = y_pred_rescaled[-prediction_length:, signal]\n", + " \n", + " # Get the true output-signal from the data-set.\n", + " signal_true = y_true[-prediction_length:, signal]\n", + "\n", + " # Make the plotting-canvas bigger.\n", + " plt.figure(figsize=(13,3))\n", + " \n", + " # Plot and compare the two signals.\n", + " plt.plot(signal_true, label='true')\n", + " plt.plot(signal_pred, label='pred')\n", + "\n", + " # Plot labels etc.\n", + " plt.ylabel(target_names[signal])\n", + " plt.legend()\n", + " plt.show()" + ] + }, + { + "cell_type": "code", + "execution_count": 283, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "#plot_last_week(train=True)\n", + "plot_last_week()" + ] + } + ], + "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": "3.6.4" + }, + "toc": { + "base_numbering": 1, + "nav_menu": {}, + "number_sections": true, + "sideBar": true, + "skip_h1_title": false, + "title_cell": "Table of Contents", + "title_sidebar": "Contents", + "toc_cell": false, + "toc_position": {}, + "toc_section_display": true, + "toc_window_display": false + } + }, + "nbformat": 4, + "nbformat_minor": 2 +} diff --git a/05-Uber_demand_forecasting/code/02 - GRU_Uber.ipynb b/05-Uber_demand_forecasting/code/02 - GRU_Uber.ipynb index fa8eb8c..d7305d9 100644 --- a/05-Uber_demand_forecasting/code/02 - GRU_Uber.ipynb +++ b/05-Uber_demand_forecasting/code/02 - GRU_Uber.ipynb @@ -2489,7 +2489,7 @@ " model.compile(loss=loss_mse_warmup, optimizer=optimizer)\n", " print(model.summary())\n", " \n", - " # Define challbacks\n", + " # Define callbacks\n", " path_checkpoint = '23_checkpoint.keras'\n", " callback_checkpoint = ModelCheckpoint(filepath=path_checkpoint,\n", " monitor='val_loss',\n", diff --git a/05-Uber_demand_forecasting/data/.DS_Store b/05-Uber_demand_forecasting/data/.DS_Store index 9d4ae572f825b070164ae053fc5d867898e9e50e..3272dcf308d2b5e8d7159bd5c5b3690de02e3ec1 100644 GIT binary patch delta 19 acmZoMXffCz$I51GuA^XRvRQ@ofe-*R;sq-J delta 19 acmZoMXffCz$I518tfOFPx><$wfe-*RwFM^t