chapter-13-etl.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Miscellaneous Munging Methods (ETL)\n",
    "\n",
    "<a href=\"../../_slides/chapter-13-slides.html\">See also the slides that summarize a portion of this content.</a>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## What do these words mean?\n",
    "\n",
    "ETL stands for \"Extract, Transform, and Load.\"  This is the standard term for all the work you may need to do with data to get it ready for actual analysis.  Before we get to make attractive visualizations or do useful analyses and produce insights, we have to get the data into a form that makes those things possible.  Think of the terms as having roughly these meanings:\n",
    " * Extract = get data from the web, a database, or wherever it's originally located (and maybe save it into a CSV file on our computer, for example)\n",
    " * Transform = manipulate the content of the data to make it more suitable for our needs (such as converting column data types, handling missing values, etc.)\n",
    " * Load = get the data into our Python script, notebook, or other analysis software (which can be an easy one-liner for small data, but is harder for big data)\n",
    "\n",
    "While ETL is an official term, the slang term is \"munging.\"  The word is well-chosen, in that it sounds a little bit awkward and a little bit gross.  Like data manipulation often is.  When most people say \"data munging,\" they're probably referring more to the \"transform\" part of ETL.\n",
    "\n",
    "I suspect people say ETL when they're speaking professionally and they say munging when they're complaining to a friend.\n",
    "\n",
    "## Why are we focusing on this?\n",
    "\n",
    "```{admonition} Big Picture - Munging/ETL is a large portion of data work\n",
    "---\n",
    "class: alert alert-primary\n",
    "---\n",
    "Many well-respected people in the data science community estimate that 70% to 80% of a data scientist's time can be spent on ETL rather than on the more interesting work of modeling, analysis, visualization, and communication.\n",
    "```\n",
    "\n",
    "While many people hear those high percentages and can't believe it, I suspect that by this point in our course, that doesn't sound at all unreasonable to you.  Just last week, our in-class exercise was to merge two datasets, which takes only two or three lines of Python code.  But the amount of work necessary to prepare the datasets for a useful merge was far greater.\n",
    "\n",
    "In *The Data Science Design Manual,* Steven Skeina has a useful chapter on ETL.  He has a humorous way of expressing the idea that ETL is a huge part of data work:\n",
    "\n",
    "> Most data scientists spend much of their time cleaning and formatting data. The rest spend most of their time complaining that there is no data available to do what they want to do.\n",
    "\n",
    "In other words, you can buckle down and do the munging you need to get the data you want, or you can sit around and get nowhere.  Those are the options.\n",
    "\n",
    "Well, okay, there is a more pleasant option.  You can advance far enough in an organization that you have data workers under you in the org chart, and you make them do the ETL and hand you the results so that you can do the interesting stuff.  But you have to put in your time as a new hire before you can rise to directing others, and even then, you'll still have to work closely with those you supervise to be sure that their munging gives you the kind of result you can use.\n",
    "\n",
    "Now, the variety of things that fall under the ETL category is truly enormous.  The reason for this is that the purpose of munging is to take ugliness and clean it up, and there are so many different types of ugliness in the world.  [When discussing tidy data](https://vita.had.co.nz/papers/tidy-data.pdf), Hadley Wickham quotes Tolstoy:\n",
    "\n",
    "> Happy families are all alike; every unhappy family is unhappy in its own way.\n",
    "\n",
    "Because every dataset is unhappy in its own way, your munging toolbelt can never be too big.  And so we can't possibly cover it all in this chapter.  Experience with datasets is the best teacher, and I intend this course to give you many experiences with new datasets.  But we will cover some key topics."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data provenance\n",
    "\n",
    "### What is provenance?\n",
    "\n",
    "If you've ever watched [Antiques Roadshow](https://www.pbs.org/wgbh/roadshow/) (or walked in while one of your grandparents was watching it), you'll know that the value of an item can be significantly impacted by its *provenance,* which means its history and origins.  If the appraiser can verify that a particular antique item was part of an important event or story in the past, or that the item is officially documented as being genuine, then this increases the item's value.\n",
    "\n",
    "The value of data is also significantly impacted by its history and origins.  If we know how the data was collected and can read about the details of that process, that will probably significantly increase its usefulness to us.\n",
    "\n",
    "For instance, imagine you get a dataset in which some numeric columns are entitled `EQ50`, `EQ51`, `EQ52`, and so on.  You would probably not be able to use the numbers in those columns for any purpose, because you don't know what they mean.  But now imagine that you find out that the data came from an economic survey that happened every quarter, and measured the GDP of various U.S. states during that quarter, in units of millions of dollars.  The organization that did the work referred to such measurements as \"Economic Quarters\" or EQs for short, and started with EQ1 in January 1987, counting upwards from there.  We can therefore figure out that EQ50 must refer to the second quarter of 2000, and so on.  Formerly useless data now has meaning and could be used.\n",
    "\n",
    "### Data vs. information\n",
    "\n",
    "```{admonition} Big Picture - Information = Data + Context\n",
    "---\n",
    "class: alert alert-primary\n",
    "---\n",
    "The difference between *data* and *information* is context.  Data is raw numbers, while information is having those numbers in a context we understand, so that the numbers have meaning.  Data provenance can be the context that turns data into information.\n",
    "```\n",
    "\n",
    "To make sense of data (that is, to have information, not just data) requires knowing something about the domain in which the data lives.  [One of the examples in the previous chapter](chapter-12-concat-and-merge.html#when-there-are-many-matches-for-some-rows) was about data from American football.  If you're not familiar with that sport, it's harder to understand the example, so I was careful to explain in the chapter the few necessary football concepts you'd need.  If your dataset comes from finance, you'll be better equipped to turn that data into information if you know something about finance.  If you're working with economic data, you'll do better if you know economics.\n",
    "\n",
    "This is where Bentley students have an advantage in data science over students from other universities.  While some schools have excellent technical educations and may cover more programming or machine learning skills than a data degree from Bentley does, every Bentley graduate has undergone an extensive training in business.  If you're planning on applying your data skills in the business world, you'll have a broader knowledge of that domain than most students from, say, an engineering school or a computer science degree.\n",
    "\n",
    "### Data dictionaries\n",
    "\n",
    "Anyone producing a dataset should take care to distribute with it a data dictionary, which is a human-readable explanation in clear language of the meaning of each column in the dataset.  We've referred very often to the home mortgage dataset in these notes; it comes with an extensive data dictionary provided by the Consumer Financial Protecion Bureau, and you can [see it online here](https://ffiec.cfpb.gov/documentation/2018/lar-data-fields/).  Since the average person doesn't know what column names like \"lei\" or \"hoepa_status\" or \"aus-4\" might mean, it's essential to be able to look them up in a data dictionary.\n",
    "\n",
    "If your employer puts you in charge of creating a dataset to be used by others, be sure that you always couple it with a document explaining the meaning of each column.  If you find a dataset you'd like to use in your own work (whether it comes from the web for your use in MA346 or it comes from your company's intranet when you have an internship or job), one of the first questions you should ask is where the data dictionary is.  Otherwise, how will you know what the data means?\n",
    "\n",
    "If a dataset doesn't come from a data dictionary, but you have personal access to the source of the data (such as another team within your company), you can organize a meeting to ask them where the data comes from and what its columns mean.  Documenting the results of such a meeting and storing it with the data in a data dictionary make that dataset more useful to everyone thereafter (and save everyone from repeating the same meeting later).  I had a meeting of exactly this type with the nonprofit organization that partnered with my graduate data science class in Fall 2019 to discuss their datasets."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Missing values\n",
    "\n",
    "This can be one of the most confusing aspects of data work for new students of data science, so let me begin by emphasizing four key points about missing values that you should always keep in mind.\n",
    "\n",
    "```{admonition} Big Picture - Summary of key points about missing values\n",
    "---\n",
    "class: alert alert-primary\n",
    "---\n",
    "1. Missing values are extremely common, and are sometimes inevitable.\n",
    "2. Sometimes missing values indicate a mistake or a problem, and sometimes they don't.\n",
    "3. Replacing missing values with actual values is called *imputation,* and there are *many* different ways to do it.\n",
    "4. Sometimes imputation is the right thing to do with missing values, but sometimes it is the wrong thing to do.\n",
    "```\n",
    "\n",
    "Let's think through the details of these important points.\n",
    "\n",
    "### Why missing values are everywhere\n",
    "\n",
    "Missing values can and do appear in almost every type of dataset.  In the home mortgage dataset, for instance, anyone who didn't fully complete the application will have some parts of their record in the database missing.  When compiling a comprehensive record of millions of mortgage applications, we simply can't expect that everyone filled out the application completely!  Missing values are inevitable.\n",
    "\n",
    "Even if you imagine a much more reliable source of data than human beings, such as a robotic sensor that's programmed to take weather readings every hour on the hour, things can still go wrong.  The sensor can fail and not collect data for a few hours until someone replaces it and reconnects it.  The people in charge of the experiment can accidentally delete or lose some data files.  The hard drive on which the data is stored can malfunction so that not all data can be recovered.  Missing values can happen anywhere.\n",
    "\n",
    "### Are missing values bad?\n",
    "\n",
    "Sometimes missing values occur in a dataset because of a problem.  Consider the examples given in the previous paragraph.  A broken sensor that fails to report data for a few hours means that something went wrong, something we wish hadn't happened, but now our data is incomplete because of that problem.  The missing values reflect that problem.\n",
    "\n",
    "But sometimes missing values are inserted into a dataset intentionally, because the creator of the dataset wants to communicate that a certain piece of data is unavailable.  For instance, in my football dataset, if the Receiver column in the Plays table has some missing entries, that means that there was no receiver involved in the play.  The missing values are communicating something intentional, sensible, and correct.  *Missing values don't always indicate a problem.*\n",
    "\n",
    "Even in the example of the failed sensor, where the missing values indicate a problem, that doesn't mean that they should be removed or filled in with actual values.  Those missing values are truthfully stating when data was not collected.  Altering them would mean that our dataset would no longer be telling the truth about its origins.  If you're sworn in on the witness stand, and you're asked who committed the robbery, and you honestly don't know the answer, the truthful thing to do is to say that you don't know!  Making up an answer is clearly a deceptive thing to do in that situation, and making up values is often a deceptive thing to do with data as well.  *Resist the urge to \"solve\" missing values by always filling them in.*  Sometimes they're telling an important truth.\n",
    "\n",
    "In fact, this is why NumPy has the built-in value `np.nan`, Python has `None`, R has `NA`, and Julia has `missing`.  These languages all recognize the legitimacy of missing values, and give you a way to express them when you need to.\n",
    "\n",
    "Notice the connection between these issues and data provenance.  If we know where the data came from and how it was obtained, we might be able to make sense of the missing values, and they can have important meaning for us, even though they're missing.\n",
    "\n",
    "### Should I ever remove missing values?\n",
    "\n",
    "**Example 1: Removing missing values**\n",
    "\n",
    "Some circumstances demand that we remove missing values.  Consider the following (real) dataset of the number of home runs hit per game in each Major League Baseball World Series in the 1990s."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "   Year    HR  #Games   HR/Game\n",
       "0  1990   6.0     4.0  1.500000\n",
       "1  1991  16.0     7.0  2.285714\n",
       "2  1992   9.0     6.0  1.500000\n",
       "3  1993  13.0     6.0  2.166667\n",
       "4  1994   NaN     NaN       NaN\n",
       "5  1995  13.0     6.0  2.166667\n",
       "6  1996   6.0     6.0  1.000000\n",
       "7  1997  15.0     7.0  2.142857\n",
       "8  1998   9.0     4.0  2.250000\n",
       "9  1999   6.0     4.0  1.500000"
      ],
      "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Year</th>\n      <th>HR</th>\n      <th>#Games</th>\n      <th>HR/Game</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>1990</td>\n      <td>6.0</td>\n      <td>4.0</td>\n      <td>1.500000</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>1991</td>\n      <td>16.0</td>\n      <td>7.0</td>\n      <td>2.285714</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>1992</td>\n      <td>9.0</td>\n      <td>6.0</td>\n      <td>1.500000</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>1993</td>\n      <td>13.0</td>\n      <td>6.0</td>\n      <td>2.166667</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>1994</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n    </tr>\n    <tr>\n      <th>5</th>\n      <td>1995</td>\n      <td>13.0</td>\n      <td>6.0</td>\n      <td>2.166667</td>\n    </tr>\n    <tr>\n      <th>6</th>\n      <td>1996</td>\n      <td>6.0</td>\n      <td>6.0</td>\n      <td>1.000000</td>\n    </tr>\n    <tr>\n      <th>7</th>\n      <td>1997</td>\n      <td>15.0</td>\n      <td>7.0</td>\n      <td>2.142857</td>\n    </tr>\n    <tr>\n      <th>8</th>\n      <td>1998</td>\n      <td>9.0</td>\n      <td>4.0</td>\n      <td>2.250000</td>\n    </tr>\n    <tr>\n      <th>9</th>\n      <td>1999</td>\n      <td>6.0</td>\n      <td>4.0</td>\n      <td>1.500000</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "metadata": {},
     "execution_count": 2
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "df = pd.DataFrame( {\n",
    "    # This data was collected by hand from pages on baseball-reference.com.\n",
    "    \"Year\" : [ 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999 ],\n",
    "    \"HR\" : [ 6, 16, 9, 13, np.nan, 13, 6, 15, 9, 6 ],\n",
    "    \"#Games\" : [ 4, 7, 6, 6, np.nan, 6, 6, 7, 4, 4 ]\n",
    "} )\n",
    "df['HR/Game'] = df['HR'] / df['#Games']\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/svg+xml": "<?xml version=\"1.0\" encoding=\"utf-8\" standalone=\"no\"?>\n<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\n  \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\n<!-- Created with matplotlib (https://matplotlib.org/) -->\n<svg height=\"277.314375pt\" version=\"1.1\" viewBox=\"0 0 385.78125 277.314375\" width=\"385.78125pt\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n <metadata>\n  <rdf:RDF xmlns:cc=\"http://creativecommons.org/ns#\" xmlns:dc=\"http://purl.org/dc/elements/1.1/\" xmlns:rdf=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#\">\n   <cc:Work>\n    <dc:type rdf:resource=\"http://purl.org/dc/dcmitype/StillImage\"/>\n    <dc:date>2021-05-14T19:08:52.245195</dc:date>\n    <dc:format>image/svg+xml</dc:format>\n    <dc:creator>\n     <cc:Agent>\n      <dc:title>Matplotlib v3.3.2, https://matplotlib.org/</dc:title>\n     </cc:Agent>\n    </dc:creator>\n   </cc:Work>\n  </rdf:RDF>\n </metadata>\n <defs>\n  <style type=\"text/css\">*{stroke-linecap:butt;stroke-linejoin:round;}</style>\n </defs>\n <g id=\"figure_1\">\n  <g id=\"patch_1\">\n   <path d=\"M 0 277.314375 \nL 385.78125 277.314375 \nL 385.78125 0 \nL 0 0 \nz\n\" style=\"fill:none;\"/>\n  </g>\n  <g id=\"axes_1\">\n   <g id=\"patch_2\">\n    <path d=\"M 43.78125 239.758125 \nL 378.58125 239.758125 \nL 378.58125 22.318125 \nL 43.78125 22.318125 \nz\n\" style=\"fill:#ffffff;\"/>\n   </g>\n   <g id=\"matplotlib.axis_1\">\n    <g id=\"xtick_1\">\n     <g id=\"line2d_1\">\n      <defs>\n       <path d=\"M 0 0 \nL 0 3.5 \n\" id=\"m23dcd0356f\" style=\"stroke:#000000;stroke-width:0.8;\"/>\n      </defs>\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"58.999432\" xlink:href=\"#m23dcd0356f\" y=\"239.758125\"/>\n      </g>\n     </g>\n     <g id=\"text_1\">\n      <!-- 1990 -->\n      <g transform=\"translate(46.274432 254.356562)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 12.40625 8.296875 \nL 28.515625 8.296875 \nL 28.515625 63.921875 \nL 10.984375 60.40625 \nL 10.984375 69.390625 \nL 28.421875 72.90625 \nL 38.28125 72.90625 \nL 38.28125 8.296875 \nL 54.390625 8.296875 \nL 54.390625 0 \nL 12.40625 0 \nz\n\" id=\"DejaVuSans-49\"/>\n        <path d=\"M 10.984375 1.515625 \nL 10.984375 10.5 \nQ 14.703125 8.734375 18.5 7.8125 \nQ 22.3125 6.890625 25.984375 6.890625 \nQ 35.75 6.890625 40.890625 13.453125 \nQ 46.046875 20.015625 46.78125 33.40625 \nQ 43.953125 29.203125 39.59375 26.953125 \nQ 35.25 24.703125 29.984375 24.703125 \nQ 19.046875 24.703125 12.671875 31.3125 \nQ 6.296875 37.9375 6.296875 49.421875 \nQ 6.296875 60.640625 12.9375 67.421875 \nQ 19.578125 74.21875 30.609375 74.21875 \nQ 43.265625 74.21875 49.921875 64.515625 \nQ 56.59375 54.828125 56.59375 36.375 \nQ 56.59375 19.140625 48.40625 8.859375 \nQ 40.234375 -1.421875 26.421875 -1.421875 \nQ 22.703125 -1.421875 18.890625 -0.6875 \nQ 15.09375 0.046875 10.984375 1.515625 \nz\nM 30.609375 32.421875 \nQ 37.25 32.421875 41.125 36.953125 \nQ 45.015625 41.5 45.015625 49.421875 \nQ 45.015625 57.28125 41.125 61.84375 \nQ 37.25 66.40625 30.609375 66.40625 \nQ 23.96875 66.40625 20.09375 61.84375 \nQ 16.21875 57.28125 16.21875 49.421875 \nQ 16.21875 41.5 20.09375 36.953125 \nQ 23.96875 32.421875 30.609375 32.421875 \nz\n\" id=\"DejaVuSans-57\"/>\n        <path d=\"M 31.78125 66.40625 \nQ 24.171875 66.40625 20.328125 58.90625 \nQ 16.5 51.421875 16.5 36.375 \nQ 16.5 21.390625 20.328125 13.890625 \nQ 24.171875 6.390625 31.78125 6.390625 \nQ 39.453125 6.390625 43.28125 13.890625 \nQ 47.125 21.390625 47.125 36.375 \nQ 47.125 51.421875 43.28125 58.90625 \nQ 39.453125 66.40625 31.78125 66.40625 \nz\nM 31.78125 74.21875 \nQ 44.046875 74.21875 50.515625 64.515625 \nQ 56.984375 54.828125 56.984375 36.375 \nQ 56.984375 17.96875 50.515625 8.265625 \nQ 44.046875 -1.421875 31.78125 -1.421875 \nQ 19.53125 -1.421875 13.0625 8.265625 \nQ 6.59375 17.96875 6.59375 36.375 \nQ 6.59375 54.828125 13.0625 64.515625 \nQ 19.53125 74.21875 31.78125 74.21875 \nz\n\" id=\"DejaVuSans-48\"/>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"190.869141\" xlink:href=\"#DejaVuSans-48\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_2\">\n     <g id=\"line2d_2\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"92.817614\" xlink:href=\"#m23dcd0356f\" y=\"239.758125\"/>\n      </g>\n     </g>\n     <g id=\"text_2\">\n      <!-- 1991 -->\n      <g transform=\"translate(80.092614 254.356562)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"190.869141\" xlink:href=\"#DejaVuSans-49\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_3\">\n     <g id=\"line2d_3\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"126.635795\" xlink:href=\"#m23dcd0356f\" y=\"239.758125\"/>\n      </g>\n     </g>\n     <g id=\"text_3\">\n      <!-- 1992 -->\n      <g transform=\"translate(113.910795 254.356562)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 19.1875 8.296875 \nL 53.609375 8.296875 \nL 53.609375 0 \nL 7.328125 0 \nL 7.328125 8.296875 \nQ 12.9375 14.109375 22.625 23.890625 \nQ 32.328125 33.6875 34.8125 36.53125 \nQ 39.546875 41.84375 41.421875 45.53125 \nQ 43.3125 49.21875 43.3125 52.78125 \nQ 43.3125 58.59375 39.234375 62.25 \nQ 35.15625 65.921875 28.609375 65.921875 \nQ 23.96875 65.921875 18.8125 64.3125 \nQ 13.671875 62.703125 7.8125 59.421875 \nL 7.8125 69.390625 \nQ 13.765625 71.78125 18.9375 73 \nQ 24.125 74.21875 28.421875 74.21875 \nQ 39.75 74.21875 46.484375 68.546875 \nQ 53.21875 62.890625 53.21875 53.421875 \nQ 53.21875 48.921875 51.53125 44.890625 \nQ 49.859375 40.875 45.40625 35.40625 \nQ 44.1875 33.984375 37.640625 27.21875 \nQ 31.109375 20.453125 19.1875 8.296875 \nz\n\" id=\"DejaVuSans-50\"/>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"190.869141\" xlink:href=\"#DejaVuSans-50\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_4\">\n     <g id=\"line2d_4\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"160.453977\" xlink:href=\"#m23dcd0356f\" y=\"239.758125\"/>\n      </g>\n     </g>\n     <g id=\"text_4\">\n      <!-- 1993 -->\n      <g transform=\"translate(147.728977 254.356562)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 40.578125 39.3125 \nQ 47.65625 37.796875 51.625 33 \nQ 55.609375 28.21875 55.609375 21.1875 \nQ 55.609375 10.40625 48.1875 4.484375 \nQ 40.765625 -1.421875 27.09375 -1.421875 \nQ 22.515625 -1.421875 17.65625 -0.515625 \nQ 12.796875 0.390625 7.625 2.203125 \nL 7.625 11.71875 \nQ 11.71875 9.328125 16.59375 8.109375 \nQ 21.484375 6.890625 26.8125 6.890625 \nQ 36.078125 6.890625 40.9375 10.546875 \nQ 45.796875 14.203125 45.796875 21.1875 \nQ 45.796875 27.640625 41.28125 31.265625 \nQ 36.765625 34.90625 28.71875 34.90625 \nL 20.21875 34.90625 \nL 20.21875 43.015625 \nL 29.109375 43.015625 \nQ 36.375 43.015625 40.234375 45.921875 \nQ 44.09375 48.828125 44.09375 54.296875 \nQ 44.09375 59.90625 40.109375 62.90625 \nQ 36.140625 65.921875 28.71875 65.921875 \nQ 24.65625 65.921875 20.015625 65.03125 \nQ 15.375 64.15625 9.8125 62.3125 \nL 9.8125 71.09375 \nQ 15.4375 72.65625 20.34375 73.4375 \nQ 25.25 74.21875 29.59375 74.21875 \nQ 40.828125 74.21875 47.359375 69.109375 \nQ 53.90625 64.015625 53.90625 55.328125 \nQ 53.90625 49.265625 50.4375 45.09375 \nQ 46.96875 40.921875 40.578125 39.3125 \nz\n\" id=\"DejaVuSans-51\"/>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"190.869141\" xlink:href=\"#DejaVuSans-51\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_5\">\n     <g id=\"line2d_5\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"194.272159\" xlink:href=\"#m23dcd0356f\" y=\"239.758125\"/>\n      </g>\n     </g>\n     <g id=\"text_5\">\n      <!-- 1994 -->\n      <g transform=\"translate(181.547159 254.356562)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 37.796875 64.3125 \nL 12.890625 25.390625 \nL 37.796875 25.390625 \nz\nM 35.203125 72.90625 \nL 47.609375 72.90625 \nL 47.609375 25.390625 \nL 58.015625 25.390625 \nL 58.015625 17.1875 \nL 47.609375 17.1875 \nL 47.609375 0 \nL 37.796875 0 \nL 37.796875 17.1875 \nL 4.890625 17.1875 \nL 4.890625 26.703125 \nz\n\" id=\"DejaVuSans-52\"/>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"190.869141\" xlink:href=\"#DejaVuSans-52\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_6\">\n     <g id=\"line2d_6\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"228.090341\" xlink:href=\"#m23dcd0356f\" y=\"239.758125\"/>\n      </g>\n     </g>\n     <g id=\"text_6\">\n      <!-- 1995 -->\n      <g transform=\"translate(215.365341 254.356562)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 10.796875 72.90625 \nL 49.515625 72.90625 \nL 49.515625 64.59375 \nL 19.828125 64.59375 \nL 19.828125 46.734375 \nQ 21.96875 47.46875 24.109375 47.828125 \nQ 26.265625 48.1875 28.421875 48.1875 \nQ 40.625 48.1875 47.75 41.5 \nQ 54.890625 34.8125 54.890625 23.390625 \nQ 54.890625 11.625 47.5625 5.09375 \nQ 40.234375 -1.421875 26.90625 -1.421875 \nQ 22.3125 -1.421875 17.546875 -0.640625 \nQ 12.796875 0.140625 7.71875 1.703125 \nL 7.71875 11.625 \nQ 12.109375 9.234375 16.796875 8.0625 \nQ 21.484375 6.890625 26.703125 6.890625 \nQ 35.15625 6.890625 40.078125 11.328125 \nQ 45.015625 15.765625 45.015625 23.390625 \nQ 45.015625 31 40.078125 35.4375 \nQ 35.15625 39.890625 26.703125 39.890625 \nQ 22.75 39.890625 18.8125 39.015625 \nQ 14.890625 38.140625 10.796875 36.28125 \nz\n\" id=\"DejaVuSans-53\"/>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"190.869141\" xlink:href=\"#DejaVuSans-53\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_7\">\n     <g id=\"line2d_7\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"261.908523\" xlink:href=\"#m23dcd0356f\" y=\"239.758125\"/>\n      </g>\n     </g>\n     <g id=\"text_7\">\n      <!-- 1996 -->\n      <g transform=\"translate(249.183523 254.356562)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 33.015625 40.375 \nQ 26.375 40.375 22.484375 35.828125 \nQ 18.609375 31.296875 18.609375 23.390625 \nQ 18.609375 15.53125 22.484375 10.953125 \nQ 26.375 6.390625 33.015625 6.390625 \nQ 39.65625 6.390625 43.53125 10.953125 \nQ 47.40625 15.53125 47.40625 23.390625 \nQ 47.40625 31.296875 43.53125 35.828125 \nQ 39.65625 40.375 33.015625 40.375 \nz\nM 52.59375 71.296875 \nL 52.59375 62.3125 \nQ 48.875 64.0625 45.09375 64.984375 \nQ 41.3125 65.921875 37.59375 65.921875 \nQ 27.828125 65.921875 22.671875 59.328125 \nQ 17.53125 52.734375 16.796875 39.40625 \nQ 19.671875 43.65625 24.015625 45.921875 \nQ 28.375 48.1875 33.59375 48.1875 \nQ 44.578125 48.1875 50.953125 41.515625 \nQ 57.328125 34.859375 57.328125 23.390625 \nQ 57.328125 12.15625 50.6875 5.359375 \nQ 44.046875 -1.421875 33.015625 -1.421875 \nQ 20.359375 -1.421875 13.671875 8.265625 \nQ 6.984375 17.96875 6.984375 36.375 \nQ 6.984375 53.65625 15.1875 63.9375 \nQ 23.390625 74.21875 37.203125 74.21875 \nQ 40.921875 74.21875 44.703125 73.484375 \nQ 48.484375 72.75 52.59375 71.296875 \nz\n\" id=\"DejaVuSans-54\"/>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"190.869141\" xlink:href=\"#DejaVuSans-54\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_8\">\n     <g id=\"line2d_8\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"295.726705\" xlink:href=\"#m23dcd0356f\" y=\"239.758125\"/>\n      </g>\n     </g>\n     <g id=\"text_8\">\n      <!-- 1997 -->\n      <g transform=\"translate(283.001705 254.356562)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 8.203125 72.90625 \nL 55.078125 72.90625 \nL 55.078125 68.703125 \nL 28.609375 0 \nL 18.3125 0 \nL 43.21875 64.59375 \nL 8.203125 64.59375 \nz\n\" id=\"DejaVuSans-55\"/>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"190.869141\" xlink:href=\"#DejaVuSans-55\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_9\">\n     <g id=\"line2d_9\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"329.544886\" xlink:href=\"#m23dcd0356f\" y=\"239.758125\"/>\n      </g>\n     </g>\n     <g id=\"text_9\">\n      <!-- 1998 -->\n      <g transform=\"translate(316.819886 254.356562)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 31.78125 34.625 \nQ 24.75 34.625 20.71875 30.859375 \nQ 16.703125 27.09375 16.703125 20.515625 \nQ 16.703125 13.921875 20.71875 10.15625 \nQ 24.75 6.390625 31.78125 6.390625 \nQ 38.8125 6.390625 42.859375 10.171875 \nQ 46.921875 13.96875 46.921875 20.515625 \nQ 46.921875 27.09375 42.890625 30.859375 \nQ 38.875 34.625 31.78125 34.625 \nz\nM 21.921875 38.8125 \nQ 15.578125 40.375 12.03125 44.71875 \nQ 8.5 49.078125 8.5 55.328125 \nQ 8.5 64.0625 14.71875 69.140625 \nQ 20.953125 74.21875 31.78125 74.21875 \nQ 42.671875 74.21875 48.875 69.140625 \nQ 55.078125 64.0625 55.078125 55.328125 \nQ 55.078125 49.078125 51.53125 44.71875 \nQ 48 40.375 41.703125 38.8125 \nQ 48.828125 37.15625 52.796875 32.3125 \nQ 56.78125 27.484375 56.78125 20.515625 \nQ 56.78125 9.90625 50.3125 4.234375 \nQ 43.84375 -1.421875 31.78125 -1.421875 \nQ 19.734375 -1.421875 13.25 4.234375 \nQ 6.78125 9.90625 6.78125 20.515625 \nQ 6.78125 27.484375 10.78125 32.3125 \nQ 14.796875 37.15625 21.921875 38.8125 \nz\nM 18.3125 54.390625 \nQ 18.3125 48.734375 21.84375 45.5625 \nQ 25.390625 42.390625 31.78125 42.390625 \nQ 38.140625 42.390625 41.71875 45.5625 \nQ 45.3125 48.734375 45.3125 54.390625 \nQ 45.3125 60.0625 41.71875 63.234375 \nQ 38.140625 66.40625 31.78125 66.40625 \nQ 25.390625 66.40625 21.84375 63.234375 \nQ 18.3125 60.0625 18.3125 54.390625 \nz\n\" id=\"DejaVuSans-56\"/>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"190.869141\" xlink:href=\"#DejaVuSans-56\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_10\">\n     <g id=\"line2d_10\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"363.363068\" xlink:href=\"#m23dcd0356f\" y=\"239.758125\"/>\n      </g>\n     </g>\n     <g id=\"text_10\">\n      <!-- 1999 -->\n      <g transform=\"translate(350.638068 254.356562)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-57\"/>\n       <use x=\"190.869141\" xlink:href=\"#DejaVuSans-57\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"text_11\">\n     <!-- Year -->\n     <g transform=\"translate(200.59375 268.034687)scale(0.1 -0.1)\">\n      <defs>\n       <path d=\"M -0.203125 72.90625 \nL 10.40625 72.90625 \nL 30.609375 42.921875 \nL 50.6875 72.90625 \nL 61.28125 72.90625 \nL 35.5 34.71875 \nL 35.5 0 \nL 25.59375 0 \nL 25.59375 34.71875 \nz\n\" id=\"DejaVuSans-89\"/>\n       <path d=\"M 56.203125 29.59375 \nL 56.203125 25.203125 \nL 14.890625 25.203125 \nQ 15.484375 15.921875 20.484375 11.0625 \nQ 25.484375 6.203125 34.421875 6.203125 \nQ 39.59375 6.203125 44.453125 7.46875 \nQ 49.3125 8.734375 54.109375 11.28125 \nL 54.109375 2.78125 \nQ 49.265625 0.734375 44.1875 -0.34375 \nQ 39.109375 -1.421875 33.890625 -1.421875 \nQ 20.796875 -1.421875 13.15625 6.1875 \nQ 5.515625 13.8125 5.515625 26.8125 \nQ 5.515625 40.234375 12.765625 48.109375 \nQ 20.015625 56 32.328125 56 \nQ 43.359375 56 49.78125 48.890625 \nQ 56.203125 41.796875 56.203125 29.59375 \nz\nM 47.21875 32.234375 \nQ 47.125 39.59375 43.09375 43.984375 \nQ 39.0625 48.390625 32.421875 48.390625 \nQ 24.90625 48.390625 20.390625 44.140625 \nQ 15.875 39.890625 15.1875 32.171875 \nz\n\" id=\"DejaVuSans-101\"/>\n       <path d=\"M 34.28125 27.484375 \nQ 23.390625 27.484375 19.1875 25 \nQ 14.984375 22.515625 14.984375 16.5 \nQ 14.984375 11.71875 18.140625 8.90625 \nQ 21.296875 6.109375 26.703125 6.109375 \nQ 34.1875 6.109375 38.703125 11.40625 \nQ 43.21875 16.703125 43.21875 25.484375 \nL 43.21875 27.484375 \nz\nM 52.203125 31.203125 \nL 52.203125 0 \nL 43.21875 0 \nL 43.21875 8.296875 \nQ 40.140625 3.328125 35.546875 0.953125 \nQ 30.953125 -1.421875 24.3125 -1.421875 \nQ 15.921875 -1.421875 10.953125 3.296875 \nQ 6 8.015625 6 15.921875 \nQ 6 25.140625 12.171875 29.828125 \nQ 18.359375 34.515625 30.609375 34.515625 \nL 43.21875 34.515625 \nL 43.21875 35.40625 \nQ 43.21875 41.609375 39.140625 45 \nQ 35.0625 48.390625 27.6875 48.390625 \nQ 23 48.390625 18.546875 47.265625 \nQ 14.109375 46.140625 10.015625 43.890625 \nL 10.015625 52.203125 \nQ 14.9375 54.109375 19.578125 55.046875 \nQ 24.21875 56 28.609375 56 \nQ 40.484375 56 46.34375 49.84375 \nQ 52.203125 43.703125 52.203125 31.203125 \nz\n\" id=\"DejaVuSans-97\"/>\n       <path d=\"M 41.109375 46.296875 \nQ 39.59375 47.171875 37.8125 47.578125 \nQ 36.03125 48 33.890625 48 \nQ 26.265625 48 22.1875 43.046875 \nQ 18.109375 38.09375 18.109375 28.8125 \nL 18.109375 0 \nL 9.078125 0 \nL 9.078125 54.6875 \nL 18.109375 54.6875 \nL 18.109375 46.1875 \nQ 20.953125 51.171875 25.484375 53.578125 \nQ 30.03125 56 36.53125 56 \nQ 37.453125 56 38.578125 55.875 \nQ 39.703125 55.765625 41.0625 55.515625 \nz\n\" id=\"DejaVuSans-114\"/>\n      </defs>\n      <use xlink:href=\"#DejaVuSans-89\"/>\n      <use x=\"47.833984\" xlink:href=\"#DejaVuSans-101\"/>\n      <use x=\"109.357422\" xlink:href=\"#DejaVuSans-97\"/>\n      <use x=\"170.636719\" xlink:href=\"#DejaVuSans-114\"/>\n     </g>\n    </g>\n   </g>\n   <g id=\"matplotlib.axis_2\">\n    <g id=\"ytick_1\">\n     <g id=\"line2d_11\">\n      <defs>\n       <path d=\"M 0 0 \nL -3.5 0 \n\" id=\"m78536e3774\" style=\"stroke:#000000;stroke-width:0.8;\"/>\n      </defs>\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"43.78125\" xlink:href=\"#m78536e3774\" y=\"239.758125\"/>\n      </g>\n     </g>\n     <g id=\"text_12\">\n      <!-- 0.0 -->\n      <g transform=\"translate(20.878125 243.557344)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 10.6875 12.40625 \nL 21 12.40625 \nL 21 0 \nL 10.6875 0 \nz\n\" id=\"DejaVuSans-46\"/>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-48\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-46\"/>\n       <use x=\"95.410156\" xlink:href=\"#DejaVuSans-48\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"ytick_2\">\n     <g id=\"line2d_12\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"43.78125\" xlink:href=\"#m78536e3774\" y=\"196.270125\"/>\n      </g>\n     </g>\n     <g id=\"text_13\">\n      <!-- 0.5 -->\n      <g transform=\"translate(20.878125 200.069344)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-48\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-46\"/>\n       <use x=\"95.410156\" xlink:href=\"#DejaVuSans-53\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"ytick_3\">\n     <g id=\"line2d_13\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"43.78125\" xlink:href=\"#m78536e3774\" y=\"152.782125\"/>\n      </g>\n     </g>\n     <g id=\"text_14\">\n      <!-- 1.0 -->\n      <g transform=\"translate(20.878125 156.581344)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-46\"/>\n       <use x=\"95.410156\" xlink:href=\"#DejaVuSans-48\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"ytick_4\">\n     <g id=\"line2d_14\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"43.78125\" xlink:href=\"#m78536e3774\" y=\"109.294125\"/>\n      </g>\n     </g>\n     <g id=\"text_15\">\n      <!-- 1.5 -->\n      <g transform=\"translate(20.878125 113.093344)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-49\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-46\"/>\n       <use x=\"95.410156\" xlink:href=\"#DejaVuSans-53\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"ytick_5\">\n     <g id=\"line2d_15\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"43.78125\" xlink:href=\"#m78536e3774\" y=\"65.806125\"/>\n      </g>\n     </g>\n     <g id=\"text_16\">\n      <!-- 2.0 -->\n      <g transform=\"translate(20.878125 69.605344)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-50\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-46\"/>\n       <use x=\"95.410156\" xlink:href=\"#DejaVuSans-48\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"ytick_6\">\n     <g id=\"line2d_16\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"43.78125\" xlink:href=\"#m78536e3774\" y=\"22.318125\"/>\n      </g>\n     </g>\n     <g id=\"text_17\">\n      <!-- 2.5 -->\n      <g transform=\"translate(20.878125 26.117344)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-50\"/>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-46\"/>\n       <use x=\"95.410156\" xlink:href=\"#DejaVuSans-53\"/>\n      </g>\n     </g>\n    </g>\n    <g id=\"text_18\">\n     <!-- HR/Game -->\n     <g transform=\"translate(14.798438 154.842031)rotate(-90)scale(0.1 -0.1)\">\n      <defs>\n       <path d=\"M 9.8125 72.90625 \nL 19.671875 72.90625 \nL 19.671875 43.015625 \nL 55.515625 43.015625 \nL 55.515625 72.90625 \nL 65.375 72.90625 \nL 65.375 0 \nL 55.515625 0 \nL 55.515625 34.71875 \nL 19.671875 34.71875 \nL 19.671875 0 \nL 9.8125 0 \nz\n\" id=\"DejaVuSans-72\"/>\n       <path d=\"M 44.390625 34.1875 \nQ 47.5625 33.109375 50.5625 29.59375 \nQ 53.5625 26.078125 56.59375 19.921875 \nL 66.609375 0 \nL 56 0 \nL 46.6875 18.703125 \nQ 43.0625 26.03125 39.671875 28.421875 \nQ 36.28125 30.8125 30.421875 30.8125 \nL 19.671875 30.8125 \nL 19.671875 0 \nL 9.8125 0 \nL 9.8125 72.90625 \nL 32.078125 72.90625 \nQ 44.578125 72.90625 50.734375 67.671875 \nQ 56.890625 62.453125 56.890625 51.90625 \nQ 56.890625 45.015625 53.6875 40.46875 \nQ 50.484375 35.9375 44.390625 34.1875 \nz\nM 19.671875 64.796875 \nL 19.671875 38.921875 \nL 32.078125 38.921875 \nQ 39.203125 38.921875 42.84375 42.21875 \nQ 46.484375 45.515625 46.484375 51.90625 \nQ 46.484375 58.296875 42.84375 61.546875 \nQ 39.203125 64.796875 32.078125 64.796875 \nz\n\" id=\"DejaVuSans-82\"/>\n       <path d=\"M 25.390625 72.90625 \nL 33.6875 72.90625 \nL 8.296875 -9.28125 \nL 0 -9.28125 \nz\n\" id=\"DejaVuSans-47\"/>\n       <path d=\"M 59.515625 10.40625 \nL 59.515625 29.984375 \nL 43.40625 29.984375 \nL 43.40625 38.09375 \nL 69.28125 38.09375 \nL 69.28125 6.78125 \nQ 63.578125 2.734375 56.6875 0.65625 \nQ 49.8125 -1.421875 42 -1.421875 \nQ 24.90625 -1.421875 15.25 8.5625 \nQ 5.609375 18.5625 5.609375 36.375 \nQ 5.609375 54.25 15.25 64.234375 \nQ 24.90625 74.21875 42 74.21875 \nQ 49.125 74.21875 55.546875 72.453125 \nQ 61.96875 70.703125 67.390625 67.28125 \nL 67.390625 56.78125 \nQ 61.921875 61.421875 55.765625 63.765625 \nQ 49.609375 66.109375 42.828125 66.109375 \nQ 29.4375 66.109375 22.71875 58.640625 \nQ 16.015625 51.171875 16.015625 36.375 \nQ 16.015625 21.625 22.71875 14.15625 \nQ 29.4375 6.6875 42.828125 6.6875 \nQ 48.046875 6.6875 52.140625 7.59375 \nQ 56.25 8.5 59.515625 10.40625 \nz\n\" id=\"DejaVuSans-71\"/>\n       <path d=\"M 52 44.1875 \nQ 55.375 50.25 60.0625 53.125 \nQ 64.75 56 71.09375 56 \nQ 79.640625 56 84.28125 50.015625 \nQ 88.921875 44.046875 88.921875 33.015625 \nL 88.921875 0 \nL 79.890625 0 \nL 79.890625 32.71875 \nQ 79.890625 40.578125 77.09375 44.375 \nQ 74.3125 48.1875 68.609375 48.1875 \nQ 61.625 48.1875 57.5625 43.546875 \nQ 53.515625 38.921875 53.515625 30.90625 \nL 53.515625 0 \nL 44.484375 0 \nL 44.484375 32.71875 \nQ 44.484375 40.625 41.703125 44.40625 \nQ 38.921875 48.1875 33.109375 48.1875 \nQ 26.21875 48.1875 22.15625 43.53125 \nQ 18.109375 38.875 18.109375 30.90625 \nL 18.109375 0 \nL 9.078125 0 \nL 9.078125 54.6875 \nL 18.109375 54.6875 \nL 18.109375 46.1875 \nQ 21.1875 51.21875 25.484375 53.609375 \nQ 29.78125 56 35.6875 56 \nQ 41.65625 56 45.828125 52.96875 \nQ 50 49.953125 52 44.1875 \nz\n\" id=\"DejaVuSans-109\"/>\n      </defs>\n      <use xlink:href=\"#DejaVuSans-72\"/>\n      <use x=\"75.195312\" xlink:href=\"#DejaVuSans-82\"/>\n      <use x=\"144.677734\" xlink:href=\"#DejaVuSans-47\"/>\n      <use x=\"178.369141\" xlink:href=\"#DejaVuSans-71\"/>\n      <use x=\"255.859375\" xlink:href=\"#DejaVuSans-97\"/>\n      <use x=\"317.138672\" xlink:href=\"#DejaVuSans-109\"/>\n      <use x=\"414.550781\" xlink:href=\"#DejaVuSans-101\"/>\n     </g>\n    </g>\n   </g>\n   <g id=\"line2d_17\">\n    <path clip-path=\"url(#p984367a34c)\" d=\"M 58.999432 109.294125 \nL 92.817614 40.955839 \nL 126.635795 109.294125 \nL 160.453977 51.310125 \nM 228.090341 51.310125 \nL 261.908523 152.782125 \nL 295.726705 53.380982 \nL 329.544886 44.062125 \nL 363.363068 109.294125 \n\" style=\"fill:none;stroke:#1f77b4;stroke-linecap:square;stroke-width:1.5;\"/>\n   </g>\n   <g id=\"patch_3\">\n    <path d=\"M 43.78125 239.758125 \nL 43.78125 22.318125 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"/>\n   </g>\n   <g id=\"patch_4\">\n    <path d=\"M 378.58125 239.758125 \nL 378.58125 22.318125 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"/>\n   </g>\n   <g id=\"patch_5\">\n    <path d=\"M 43.78125 239.758125 \nL 378.58125 239.758125 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"/>\n   </g>\n   <g id=\"patch_6\">\n    <path d=\"M 43.78125 22.318125 \nL 378.58125 22.318125 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"/>\n   </g>\n   <g id=\"text_19\">\n    <!-- Home Runs per Games in World Series -->\n    <g transform=\"translate(95.320313 16.318125)scale(0.12 -0.12)\">\n     <defs>\n      <path d=\"M 30.609375 48.390625 \nQ 23.390625 48.390625 19.1875 42.75 \nQ 14.984375 37.109375 14.984375 27.296875 \nQ 14.984375 17.484375 19.15625 11.84375 \nQ 23.34375 6.203125 30.609375 6.203125 \nQ 37.796875 6.203125 41.984375 11.859375 \nQ 46.1875 17.53125 46.1875 27.296875 \nQ 46.1875 37.015625 41.984375 42.703125 \nQ 37.796875 48.390625 30.609375 48.390625 \nz\nM 30.609375 56 \nQ 42.328125 56 49.015625 48.375 \nQ 55.71875 40.765625 55.71875 27.296875 \nQ 55.71875 13.875 49.015625 6.21875 \nQ 42.328125 -1.421875 30.609375 -1.421875 \nQ 18.84375 -1.421875 12.171875 6.21875 \nQ 5.515625 13.875 5.515625 27.296875 \nQ 5.515625 40.765625 12.171875 48.375 \nQ 18.84375 56 30.609375 56 \nz\n\" id=\"DejaVuSans-111\"/>\n      <path id=\"DejaVuSans-32\"/>\n      <path d=\"M 8.5 21.578125 \nL 8.5 54.6875 \nL 17.484375 54.6875 \nL 17.484375 21.921875 \nQ 17.484375 14.15625 20.5 10.265625 \nQ 23.53125 6.390625 29.59375 6.390625 \nQ 36.859375 6.390625 41.078125 11.03125 \nQ 45.3125 15.671875 45.3125 23.6875 \nL 45.3125 54.6875 \nL 54.296875 54.6875 \nL 54.296875 0 \nL 45.3125 0 \nL 45.3125 8.40625 \nQ 42.046875 3.421875 37.71875 1 \nQ 33.40625 -1.421875 27.6875 -1.421875 \nQ 18.265625 -1.421875 13.375 4.4375 \nQ 8.5 10.296875 8.5 21.578125 \nz\nM 31.109375 56 \nz\n\" id=\"DejaVuSans-117\"/>\n      <path d=\"M 54.890625 33.015625 \nL 54.890625 0 \nL 45.90625 0 \nL 45.90625 32.71875 \nQ 45.90625 40.484375 42.875 44.328125 \nQ 39.84375 48.1875 33.796875 48.1875 \nQ 26.515625 48.1875 22.3125 43.546875 \nQ 18.109375 38.921875 18.109375 30.90625 \nL 18.109375 0 \nL 9.078125 0 \nL 9.078125 54.6875 \nL 18.109375 54.6875 \nL 18.109375 46.1875 \nQ 21.34375 51.125 25.703125 53.5625 \nQ 30.078125 56 35.796875 56 \nQ 45.21875 56 50.046875 50.171875 \nQ 54.890625 44.34375 54.890625 33.015625 \nz\n\" id=\"DejaVuSans-110\"/>\n      <path d=\"M 44.28125 53.078125 \nL 44.28125 44.578125 \nQ 40.484375 46.53125 36.375 47.5 \nQ 32.28125 48.484375 27.875 48.484375 \nQ 21.1875 48.484375 17.84375 46.4375 \nQ 14.5 44.390625 14.5 40.28125 \nQ 14.5 37.15625 16.890625 35.375 \nQ 19.28125 33.59375 26.515625 31.984375 \nL 29.59375 31.296875 \nQ 39.15625 29.25 43.1875 25.515625 \nQ 47.21875 21.78125 47.21875 15.09375 \nQ 47.21875 7.46875 41.1875 3.015625 \nQ 35.15625 -1.421875 24.609375 -1.421875 \nQ 20.21875 -1.421875 15.453125 -0.5625 \nQ 10.6875 0.296875 5.421875 2 \nL 5.421875 11.28125 \nQ 10.40625 8.6875 15.234375 7.390625 \nQ 20.0625 6.109375 24.8125 6.109375 \nQ 31.15625 6.109375 34.5625 8.28125 \nQ 37.984375 10.453125 37.984375 14.40625 \nQ 37.984375 18.0625 35.515625 20.015625 \nQ 33.0625 21.96875 24.703125 23.78125 \nL 21.578125 24.515625 \nQ 13.234375 26.265625 9.515625 29.90625 \nQ 5.8125 33.546875 5.8125 39.890625 \nQ 5.8125 47.609375 11.28125 51.796875 \nQ 16.75 56 26.8125 56 \nQ 31.78125 56 36.171875 55.265625 \nQ 40.578125 54.546875 44.28125 53.078125 \nz\n\" id=\"DejaVuSans-115\"/>\n      <path d=\"M 18.109375 8.203125 \nL 18.109375 -20.796875 \nL 9.078125 -20.796875 \nL 9.078125 54.6875 \nL 18.109375 54.6875 \nL 18.109375 46.390625 \nQ 20.953125 51.265625 25.265625 53.625 \nQ 29.59375 56 35.59375 56 \nQ 45.5625 56 51.78125 48.09375 \nQ 58.015625 40.1875 58.015625 27.296875 \nQ 58.015625 14.40625 51.78125 6.484375 \nQ 45.5625 -1.421875 35.59375 -1.421875 \nQ 29.59375 -1.421875 25.265625 0.953125 \nQ 20.953125 3.328125 18.109375 8.203125 \nz\nM 48.6875 27.296875 \nQ 48.6875 37.203125 44.609375 42.84375 \nQ 40.53125 48.484375 33.40625 48.484375 \nQ 26.265625 48.484375 22.1875 42.84375 \nQ 18.109375 37.203125 18.109375 27.296875 \nQ 18.109375 17.390625 22.1875 11.75 \nQ 26.265625 6.109375 33.40625 6.109375 \nQ 40.53125 6.109375 44.609375 11.75 \nQ 48.6875 17.390625 48.6875 27.296875 \nz\n\" id=\"DejaVuSans-112\"/>\n      <path d=\"M 9.421875 54.6875 \nL 18.40625 54.6875 \nL 18.40625 0 \nL 9.421875 0 \nz\nM 9.421875 75.984375 \nL 18.40625 75.984375 \nL 18.40625 64.59375 \nL 9.421875 64.59375 \nz\n\" id=\"DejaVuSans-105\"/>\n      <path d=\"M 3.328125 72.90625 \nL 13.28125 72.90625 \nL 28.609375 11.28125 \nL 43.890625 72.90625 \nL 54.984375 72.90625 \nL 70.3125 11.28125 \nL 85.59375 72.90625 \nL 95.609375 72.90625 \nL 77.296875 0 \nL 64.890625 0 \nL 49.515625 63.28125 \nL 33.984375 0 \nL 21.578125 0 \nz\n\" id=\"DejaVuSans-87\"/>\n      <path d=\"M 9.421875 75.984375 \nL 18.40625 75.984375 \nL 18.40625 0 \nL 9.421875 0 \nz\n\" id=\"DejaVuSans-108\"/>\n      <path d=\"M 45.40625 46.390625 \nL 45.40625 75.984375 \nL 54.390625 75.984375 \nL 54.390625 0 \nL 45.40625 0 \nL 45.40625 8.203125 \nQ 42.578125 3.328125 38.25 0.953125 \nQ 33.9375 -1.421875 27.875 -1.421875 \nQ 17.96875 -1.421875 11.734375 6.484375 \nQ 5.515625 14.40625 5.515625 27.296875 \nQ 5.515625 40.1875 11.734375 48.09375 \nQ 17.96875 56 27.875 56 \nQ 33.9375 56 38.25 53.625 \nQ 42.578125 51.265625 45.40625 46.390625 \nz\nM 14.796875 27.296875 \nQ 14.796875 17.390625 18.875 11.75 \nQ 22.953125 6.109375 30.078125 6.109375 \nQ 37.203125 6.109375 41.296875 11.75 \nQ 45.40625 17.390625 45.40625 27.296875 \nQ 45.40625 37.203125 41.296875 42.84375 \nQ 37.203125 48.484375 30.078125 48.484375 \nQ 22.953125 48.484375 18.875 42.84375 \nQ 14.796875 37.203125 14.796875 27.296875 \nz\n\" id=\"DejaVuSans-100\"/>\n      <path d=\"M 53.515625 70.515625 \nL 53.515625 60.890625 \nQ 47.90625 63.578125 42.921875 64.890625 \nQ 37.9375 66.21875 33.296875 66.21875 \nQ 25.25 66.21875 20.875 63.09375 \nQ 16.5 59.96875 16.5 54.203125 \nQ 16.5 49.359375 19.40625 46.890625 \nQ 22.3125 44.4375 30.421875 42.921875 \nL 36.375 41.703125 \nQ 47.40625 39.59375 52.65625 34.296875 \nQ 57.90625 29 57.90625 20.125 \nQ 57.90625 9.515625 50.796875 4.046875 \nQ 43.703125 -1.421875 29.984375 -1.421875 \nQ 24.8125 -1.421875 18.96875 -0.25 \nQ 13.140625 0.921875 6.890625 3.21875 \nL 6.890625 13.375 \nQ 12.890625 10.015625 18.65625 8.296875 \nQ 24.421875 6.59375 29.984375 6.59375 \nQ 38.421875 6.59375 43.015625 9.90625 \nQ 47.609375 13.234375 47.609375 19.390625 \nQ 47.609375 24.75 44.3125 27.78125 \nQ 41.015625 30.8125 33.5 32.328125 \nL 27.484375 33.5 \nQ 16.453125 35.6875 11.515625 40.375 \nQ 6.59375 45.0625 6.59375 53.421875 \nQ 6.59375 63.09375 13.40625 68.65625 \nQ 20.21875 74.21875 32.171875 74.21875 \nQ 37.3125 74.21875 42.625 73.28125 \nQ 47.953125 72.359375 53.515625 70.515625 \nz\n\" id=\"DejaVuSans-83\"/>\n     </defs>\n     <use xlink:href=\"#DejaVuSans-72\"/>\n     <use x=\"75.195312\" xlink:href=\"#DejaVuSans-111\"/>\n     <use x=\"136.376953\" xlink:href=\"#DejaVuSans-109\"/>\n     <use x=\"233.789062\" xlink:href=\"#DejaVuSans-101\"/>\n     <use x=\"295.3125\" xlink:href=\"#DejaVuSans-32\"/>\n     <use x=\"327.099609\" xlink:href=\"#DejaVuSans-82\"/>\n     <use x=\"392.082031\" xlink:href=\"#DejaVuSans-117\"/>\n     <use x=\"455.460938\" xlink:href=\"#DejaVuSans-110\"/>\n     <use x=\"518.839844\" xlink:href=\"#DejaVuSans-115\"/>\n     <use x=\"570.939453\" xlink:href=\"#DejaVuSans-32\"/>\n     <use x=\"602.726562\" xlink:href=\"#DejaVuSans-112\"/>\n     <use x=\"666.203125\" xlink:href=\"#DejaVuSans-101\"/>\n     <use x=\"727.726562\" xlink:href=\"#DejaVuSans-114\"/>\n     <use x=\"768.839844\" xlink:href=\"#DejaVuSans-32\"/>\n     <use x=\"800.626953\" xlink:href=\"#DejaVuSans-71\"/>\n     <use x=\"878.117188\" xlink:href=\"#DejaVuSans-97\"/>\n     <use x=\"939.396484\" xlink:href=\"#DejaVuSans-109\"/>\n     <use x=\"1036.808594\" xlink:href=\"#DejaVuSans-101\"/>\n     <use x=\"1098.332031\" xlink:href=\"#DejaVuSans-115\"/>\n     <use x=\"1150.431641\" xlink:href=\"#DejaVuSans-32\"/>\n     <use x=\"1182.21875\" xlink:href=\"#DejaVuSans-105\"/>\n     <use x=\"1210.001953\" xlink:href=\"#DejaVuSans-110\"/>\n     <use x=\"1273.380859\" xlink:href=\"#DejaVuSans-32\"/>\n     <use x=\"1305.167969\" xlink:href=\"#DejaVuSans-87\"/>\n     <use x=\"1398.169922\" xlink:href=\"#DejaVuSans-111\"/>\n     <use x=\"1459.351562\" xlink:href=\"#DejaVuSans-114\"/>\n     <use x=\"1500.464844\" xlink:href=\"#DejaVuSans-108\"/>\n     <use x=\"1528.248047\" xlink:href=\"#DejaVuSans-100\"/>\n     <use x=\"1591.724609\" xlink:href=\"#DejaVuSans-32\"/>\n     <use x=\"1623.511719\" xlink:href=\"#DejaVuSans-83\"/>\n     <use x=\"1686.988281\" xlink:href=\"#DejaVuSans-101\"/>\n     <use x=\"1748.511719\" xlink:href=\"#DejaVuSans-114\"/>\n     <use x=\"1789.625\" xlink:href=\"#DejaVuSans-105\"/>\n     <use x=\"1817.408203\" xlink:href=\"#DejaVuSans-101\"/>\n     <use x=\"1878.931641\" xlink:href=\"#DejaVuSans-115\"/>\n    </g>\n   </g>\n  </g>\n </g>\n <defs>\n  <clipPath id=\"p984367a34c\">\n   <rect height=\"217.44\" width=\"334.8\" x=\"43.78125\" y=\"22.318125\"/>\n  </clipPath>\n </defs>\n</svg>\n",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     }
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.plot( df['Year'], df['HR/Game'] )\n",
    "plt.title( 'Home Runs per Games in World Series' )\n",
    "plt.xticks( df['Year'] )\n",
    "plt.xlabel( 'Year' )\n",
    "plt.ylabel( 'HR/Game' )\n",
    "plt.ylim( 0, 2.5 )\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Assume you were trying to show that this number was not going convincingly up or down throughout the 1990s (a made-up research question just as an example).  You're considering fitting a linear model to the data and showing that its slope is close to zero (perhaps even not statistically significantly different from zero).  Let's try."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "LinregressResult(slope=nan, intercept=nan, rvalue=nan, pvalue=nan, stderr=nan)"
      ]
     },
     "metadata": {},
     "execution_count": 4
    }
   ],
   "source": [
    "import scipy.stats as stats\n",
    "stats.linregress( df['Year'], df['HR/Game'] )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This has clearly failed, giving us all missing values in our linear model.  The reason, no doubt, is the missing value in our data.  (There was no World Series in 1994 due to a players' strike.)\n",
    "\n",
    "So in this case, the missing values are clearly causing a problem with what we want to do with the data.  And since we can fit a linear model to the data that remains, it would be perfectly acceptable to drop the one row that has missing values and proceed with the nine rows that remain.  *This is a case in which removing the missing values makes sense.*\n",
    "\n",
    "But we do not remove them from the original dataset; we simply don't include them in the data used to create the linear model.  The original dataset stays intact."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "LinregressResult(slope=-0.001238738738738731, intercept=4.305389317889302, rvalue=-0.008554596649575391, pvalue=0.9825737815654227, stderr=0.05472871741068167)"
      ]
     },
     "metadata": {},
     "execution_count": 5
    }
   ],
   "source": [
    "df_without_94 = df.dropna()\n",
    "stats.linregress( df_without_94['Year'], df_without_94['HR/Game'] )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we have an actual linear model.  (We're not going to analyze it here; that wasn't the point of this example.)\n",
    "\n",
    "**Example 2:  Not removing missing values**\n",
    "\n",
    "Let's say you're working for a small-but-growing automobile sales organization.  They've just opened their second location and they've realized that their growth has far outpaced their record-keeping.  They've got some spreadsheets about sales and commissions for their various employees, but it's not comprehensive because they haven't been organized about record-keeping in the past.  They've asked you to organize it into a database.\n",
    "\n",
    "Let's say you realize the data isn't that huge, so you can probably fit it in one spreadsheet.  You begin by creating a private Google Sheet and sharing the link with all the sales managers, asking them to paste in all the historic data on which they have records, to create a shared dataset that's as comprehensive as possible.  You start with columns for month, employee, manager, number of sales, commission, and others.  When the task is done, you notice that many rows have missing values for the number of sales and commission columns.  The managers knew the employees were working there that month, but they'd lost the relevant historical data in the intervening years.\n",
    "\n",
    "If you were to remove those rows from the dataset, it could make it seem as if the employee was not a part of the company or team at the time.  Thus even though those rows contain missing values, they are still communicating other important information.  In this case, you would decide not to remove the rows, even though they won't contribute much to any later analysis.\n",
    "\n",
    "Any decision like this made when constructing a dataset should be documented in its data dictionary."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Example 3: Actually adding missing values**\n",
    "\n",
    "In the home mortgage dataset with which we're very familiar, some columns (such as interest rate) contain mostly numerical data, but occasionally the word Exempt appears instead of a number.  This makes it impossible to do any computations on such columns, such as `df['interest_rate'].mean()`, because the column is text, not numeric.\n",
    "\n",
    "In this case, it can be valuable to replace the word Exempt with the actual missing value `np.nan` throughout the column, so that it can then be converted to type `float`.  In doing so, you should carefully document that all Exempt entries have become missing values, in order to facilitate analysis.  *This is a situation in which missing values are actually intentionally added!*\n",
    "\n",
    "If you needed to track which rows had originally been Exempt, you could retain the original interest rate column for reference, creating a new one as you do the replacement.  Alternately, you could create a new column that records simply a single boolean value for \"interest rate exempt\" so that you can tell missing values from Exempt values.\n",
    "\n",
    "Elsewhere in the same mortgage dataset, we find cases in which numbers like 999 were used for applicants' ages.  Clearly these are not correct values, and should be treated as a lack of data, rather than legitimate data.  Consider the alternatives for how to handle them:\n",
    "\n",
    "| If we leave numbers like 999 in the data | If we replace them with missing values |\n",
    "|------------------------------------------|----------------------------------------|\n",
    "| Statistics about age, like mean, median, etc., will be very wrong | Statistics about age will be much more accurate |\n",
    "| The number of missing values in the dataset will be very small | The number of missing values in the data will be much more accurate |"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### When I need to remove missing values, how do I?\n",
    "\n",
    "Removing missing values is called *data imputation,* which is simply the technical word for filling in values where there were none.  Imputation is an enormous area of statistics to which we cannot do justice in this chapter, but let's see why sometimes imputing values is essential.\n",
    "\n",
    "In the baseball example above, we saw that some model-fitting procedures can't work with missing values, and we need to remove them from consideration.  Now let's assume we were fitting some (more complex) model to the property values in the mortgage dataset.  If we need to drop any row in which the property value is missing, how might that cause problems?\n",
    "\n",
    "The question takes us right back to data provenance:  *Why* is the property value missing on the mortgage application?  Let's say we investigate and find that this is usually because the application was not completed by the potential borrower.  The question then arises: Are all borrowers equally likely to quit an application half way through?  If they are, then perhaps dropping such rows from the data is an acceptable move.\n",
    "\n",
    "But the government publishes the data to help combat discrimination in lending.  What if we were to look at the proportion of incomplete applications and find that it's much higher for certain ethnic groups, especially in certain locations?  Perhaps they're not completing the application because they're facing discrimination in the process and don't have the energy or ability to fight it.  If that's the case, then dropping rows with missing property values will *significantly reduce the representation of those ethnic groups in our data.*  Our model will unintentionally favor the other ethnic groups.  Not only will it make bad predictions (so we've done our data work wrong) but it will help to further the discrimination the dataset was trying to prevent (so we've made an ethical mistake as well)!\n",
    "\n",
    "So if we find that the missing values are not spread evenly across groups within our data, we can't in good conscience drop those rows.  Instead, we have to find some way to insert realistic or feasible values in place of the missing values.  Here are a few common ways to do so:\n",
    " * **Mean substitution** - Replace each missing property value with the mean property value across all rows.\n",
    " * **Model-based substitution** - Create a simple model that predicts property values based on other things, such as zip code, and use it to fill in each missing value.\n",
    " * **Random imputation** - Replace each missing property value with a randomly chosen property value from elsewhere in the dataset, or randomly chosen from other similar records (e.g., in the same state, or the same race, or the same income bracket, etc.).\n",
    "\n",
    "Again, many statistical concerns arise when doing imputation that we cannot cover in this short chapter of notes.  This is merely an introduction to the fact that this practice is an important one."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## All the other munging things\n",
    "\n",
    "As I said at the outset, it's not possible to cover everything you might need to do with data.  But here are a few essentials to keep in mind.\n",
    "\n",
    "When using data, keep in mind the units on every number, in terms as precise as you possibly can.  You can insert these units as comments in your code.  There are famous stories of [tens of millions of dollars lost in spacecraft](https://en.wikipedia.org/wiki/Mars_Climate_Orbiter#Cause_of_failure) when units were not checked correctly in computer code, so these tiny details are not unimportant!\n",
    "\n",
    "In *The Data Science Design Manual* quoted earlier, the author suggests several types of unit discrepencies to pay attention to.\n",
    " * differing standards of measurement, such as pounds vs. kilograms, or USD vs. GBP\n",
    " * the time value of money, such as USD in January 2017 vs. USD in February 2017\n",
    " * fluctuations in value, such as the price of gold at noon today vs. at 1pm today\n",
    " * discrepencies in time zones, such as the price of gold at noon today in London vs. noon today in New York\n",
    " * discrepencies in the units themselves, such as \"shares of stock\" before and after a stock split\n",
    "\n",
    "Another common units error to be aware of is the difference between percentages and proportions.  For instance, 15% is equal to the proportion 0.15.  When reporting such a value to a human reader, such as in a table of results, the percent is typically the more user-friendly choice.  When using such a value in a computation, such as multiplying to apply a percentage or proportion to a total quantity, the only correct choice is the proportion.  That is, 15% of 200 people is not $15\\times200=3000$, but $0.15\\times200=30$.\n",
    "\n",
    "Comments in code to track units can help with discrepencies like these.  See the code below that takes care with units as we adjust movie revenues for inflation in the following dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "               Title  Year  Opening Weekend (M$)\n",
       "0  Avengers: Endgame  2019               357.115\n",
       "1      The Lion King  2019               191.771\n",
       "2   The Hunger Games  2012               152.536\n",
       "3       Finding Dory  2016               135.060"
      ],
      "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Title</th>\n      <th>Year</th>\n      <th>Opening Weekend (M$)</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>Avengers: Endgame</td>\n      <td>2019</td>\n      <td>357.115</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>The Lion King</td>\n      <td>2019</td>\n      <td>191.771</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>The Hunger Games</td>\n      <td>2012</td>\n      <td>152.536</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>Finding Dory</td>\n      <td>2016</td>\n      <td>135.060</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "metadata": {},
     "execution_count": 6
    }
   ],
   "source": [
    "df_films = pd.DataFrame( {\n",
    "    'Title' : [ 'Avengers: Endgame', 'The Lion King', 'The Hunger Games', 'Finding Dory' ],\n",
    "    'Year' : [ 2019, 2019, 2012, 2016 ],\n",
    "    'Opening Weekend (M$)' : [ 357.115, 191.771, 152.536, 135.060 ]\n",
    "} )\n",
    "df_films"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "               Title  Year  Opening Weekend (M$)  Years since film  \\\n",
       "0  Avengers: Endgame  2019               357.115                 1   \n",
       "1      The Lion King  2019               191.771                 1   \n",
       "2   The Hunger Games  2012               152.536                 8   \n",
       "3       Finding Dory  2016               135.060                 4   \n",
       "\n",
       "   Inflation factor  Opening Weekend (M$2020)  \n",
       "0          1.030000                367.828450  \n",
       "1          1.030000                197.524130  \n",
       "2          1.266770                193.228041  \n",
       "3          1.125509                152.011220  "
      ],
      "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Title</th>\n      <th>Year</th>\n      <th>Opening Weekend (M$)</th>\n      <th>Years since film</th>\n      <th>Inflation factor</th>\n      <th>Opening Weekend (M$2020)</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>Avengers: Endgame</td>\n      <td>2019</td>\n      <td>357.115</td>\n      <td>1</td>\n      <td>1.030000</td>\n      <td>367.828450</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>The Lion King</td>\n      <td>2019</td>\n      <td>191.771</td>\n      <td>1</td>\n      <td>1.030000</td>\n      <td>197.524130</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>The Hunger Games</td>\n      <td>2012</td>\n      <td>152.536</td>\n      <td>8</td>\n      <td>1.266770</td>\n      <td>193.228041</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>Finding Dory</td>\n      <td>2016</td>\n      <td>135.060</td>\n      <td>4</td>\n      <td>1.125509</td>\n      <td>152.011220</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "metadata": {},
     "execution_count": 7
    }
   ],
   "source": [
    "avg_annual_inflation = 3                                # An approximate percentage\n",
    "inflation_factor = 1 + avg_annual_inflation/100         # Useful as an annual multiplier\n",
    "df_films['Years since film'] = 2020 - df_films['Year']  # Number of years elapsed\n",
    "df_films['Inflation factor'] = inflation_factor ** df_films['Years since film']\n",
    "                                                        # Multiplier to apply inflation\n",
    "df_films['Opening Weekend (M$2020)'] = df_films['Opening Weekend (M$)'] \\\n",
    "    * df_films['Inflation factor']                      # Convert all to $ millions in 2020\n",
    "df_films"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Before we finish discussing ETL, we should talk about file formats, which are a crucial part of the whole process."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Reading data files\n",
    "\n",
    "As you know from [the DataCamp assignment that corresponds to this chapter](big-cheat-sheet.html#before-day-8), there are many ways to read data into pandas.  Since you've learned some of the technical details from DataCamp, let's look at the relative pros and cons of each file format here, and add a few pieces of advice that didn't appear in the DataCamp lessons.  We start with the easiest file formats and work our way up.\n",
    "\n",
    "### Easy formats to read: CSV and TSV\n",
    "\n",
    "We've been using `pd.read_csv()` for ages, so there is no surprise here, and you've had to deal with its `encoding` parameter in the past as well.  It has tons of optional parameters, but the one introduced in the latest DataCamp lessons was `sep`, useful for reading TSV (tab-separated values) files, by choosing `sep=\"\\t\"`.\n",
    "\n",
    "One piece of advice I'll add to what DataCamp taught:  If you find the URL of a CSV file on the web, you can include that URL as the input parameter to `pd.read_csv()`, and it will download and read the file for you in one shot, without your having to manually download the file.\n",
    " * Pro: It automatically gets the latest version of the file every time you run your code.\n",
    " * Con: It accesses the Internet (which can sometimes be slow) every time you run your code.\n",
    " * Con: If the file is removed from the web, your code no longer functions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "       STATION     STATION_NAME  ELEVATION  LATITUDE  LONGITUDE  \\\n",
       "0  COOP:310301  ASHEVILLE NC US      682.1   35.5954   -82.5568   \n",
       "1  COOP:310301  ASHEVILLE NC US      682.1   35.5954   -82.5568   \n",
       "2  COOP:310301  ASHEVILLE NC US      682.1   35.5954   -82.5568   \n",
       "\n",
       "             DATE   HPCP Measurement Flag Quality Flag  \n",
       "0  20100101 00:00  99999                ]               \n",
       "1  20100101 01:00      0                g               \n",
       "2  20100102 06:00      1                                "
      ],
      "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>STATION</th>\n      <th>STATION_NAME</th>\n      <th>ELEVATION</th>\n      <th>LATITUDE</th>\n      <th>LONGITUDE</th>\n      <th>DATE</th>\n      <th>HPCP</th>\n      <th>Measurement Flag</th>\n      <th>Quality Flag</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>COOP:310301</td>\n      <td>ASHEVILLE NC US</td>\n      <td>682.1</td>\n      <td>35.5954</td>\n      <td>-82.5568</td>\n      <td>20100101 00:00</td>\n      <td>99999</td>\n      <td>]</td>\n      <td></td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>COOP:310301</td>\n      <td>ASHEVILLE NC US</td>\n      <td>682.1</td>\n      <td>35.5954</td>\n      <td>-82.5568</td>\n      <td>20100101 01:00</td>\n      <td>0</td>\n      <td>g</td>\n      <td></td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>COOP:310301</td>\n      <td>ASHEVILLE NC US</td>\n      <td>682.1</td>\n      <td>35.5954</td>\n      <td>-82.5568</td>\n      <td>20100102 06:00</td>\n      <td>1</td>\n      <td></td>\n      <td></td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "metadata": {},
     "execution_count": 8
    }
   ],
   "source": [
    "# Providing a URL directly to pd.read_csv():\n",
    "pd.read_csv( 'https://www1.ncdc.noaa.gov/pub/data/cdo/samples/PRECIP_HLY_sample_csv.csv' )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Pretty easy format to read: XLSX\n",
    "\n",
    "The `pd.read_excel()` function is nearly as easy to use as `pd.read_csv()`, with a few exceptions documented below.  You can give this function a URL also, if there's a publicly accessible Excel file on the web you want to download.  The same pros and cons apply when providing a URL to `pd.read_excel()` as they do for `pd.read_csv()`, as discussed above.\n",
    "\n",
    " 1. If you're running Python on a cloud service, you'll need the `openpyxl` module to be installed to add Excel support to pandas.  (You can tell if it's not when a `pd.read_excel()` call fails with an error about the missing module, or perhaps about the missing `xlrd` module, which is related.)  If you're on your local computer with an Anaconda installation, you may already have this module.  Otherwise, you need to run `pip install openpyxl` to add it.\n",
    " 2. You need to remember that this function returns a Python list of DataFrames, unless you choose one specific sheet, with `sheet_name='Name'` or choose one by index, with `sheet_name=0`, for example.\n",
    " 3. Excel spreadsheets may not have the data in the top left, so parameters like `usecols` and `skiprows` are often needed.  See [the official documentation](https://pandas.pydata.org/docs/reference/api/pandas.read_excel.html) for details on those parameters.\n",
    "\n",
    "### Easy format to read with occasional problems: HTML\n",
    "\n",
    "Pandas can often automatically extract tables from web pages.  Simply call `pd.read_html()` and give it the URL of the page containing the table or tables.  It has the same output type as `pd.read_excel()` does: a Python list of pandas DataFrames.  See the pros and cons listed under `pd.read_csv()` for providing live web URLs when reading data.\n",
    "\n",
    "Furthermore, depending on the quality of the web site, this function may or may not do its job.  If the HTML page is not structured particularly cleanly, I've had `pd.read_html()` fail to find one or more of the tables.  I've had to instead write code that downloads the HTML code, splits it wherever a `<table...>` tag begins, and extract the tables from those pieces with `pd.read_html()`.  This is annoying, but occasionally necessary.\n",
    "\n",
    "Note that if you don't need to get live data from the web, but are content with downloading the data once at the start of your project, there are many ways to extract tables from web pages.  You can often select the table and copy-paste into Excel, although that sometimes brings along undesired formatting that can cause problems.  There are Google Chrome extensions that specialize in extracting tables from web pages to make them easier to paste cleanly into Excel.\n",
    "\n",
    "### Not an easy format to read: JSON\n",
    "\n",
    "Although this format is not easy, it is powerful, and this is why it's very prevalent on the web.  It can represent a huge variety of different types of data, not just tabular data.  It is flexible enough to represent tabular data in a variety of ways, but also hierarchical data of any kind.  Due to its complexity, we will not fully review this here; refer to [the appropriate section of our course's coding cheat sheet](big-cheat-sheet.html#chapter-4-importing-json-data-and-working-with-apis) for some information, or [the corresponding DataCamp course](https://learn.datacamp.com/courses/streamlined-data-ingestion-with-pandas).\n",
    "\n",
    "### Not an easy source to read: SQL\n",
    "\n",
    "Rather than dive into the enormous topic of SQL databases here, I will suggest two ways that you can learn more:\n",
    " 1. Your next (and final) DataCamp assignment, for next week, will do some introductory coverage of this content.\n",
    " 2. Bentley has an entire course on SQL databases, CS350, which I recommend.\n",
    "\n",
    "Now let's consider which file format to use when you need to create a file rather than read one."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Writing data files\n",
    "\n",
    "As with all types of communication, it's essential to consider your audience when choosing a file type.  Who will use your file?\n",
    "\n",
    "### For a nontechnical audience, create an Excel file.\n",
    "\n",
    "If sharing your data with non-technical people, they will want to simply double-click the file and see its contents.  The easiest way to ensure this happens is to create an Excel file.  (To make it even easier, you can upload the file to SharePoint or Google Sheets and send only the link.  This is especially valuable if you suspect the recipient doesn't have Excel installed or might be viewing your email on a mobile device.)\n",
    "\n",
    "Just as when reading Excel files, you must have the `openpyxl` module installed; [see above for details](#pretty-easy-format-to-read-xlsx).  If you want to create an Excel file with just one sheet in it, you can make a single call to `df.to_excel()`.\n",
    "\n",
    "```python\n",
    "df_films.to_excel( 'Opening Weekends.xlsx' )\n",
    "```\n",
    "\n",
    "If you want to put several DataFrames into one Excel file, as different sheets in the workbook, then you need to get a little more fancy.  The indentation in the following code is essential (as always with Python).\n",
    "\n",
    "```python\n",
    "with pd.ExcelWriter( 'Two Things.xlsx' ) as writer:        # Open the file.\n",
    "    df.to_excel( writer, sheet_name='World Series Data' )  # Write one sheet.\n",
    "    df_films.to_excel( writer, sheet_name='Film Data' )    # Write the other.\n",
    "```\n",
    "\n",
    "Python's `with` statement lets you create a resource (in this case a new, open file) and Python will automatically close it up for you when you're done using it.  At the end of the two indented lines, Python will close the file, so that other applications can open it.\n",
    "\n",
    "For more details, see [the documentation for `df.to_excel()`](https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.to_excel.html).\n",
    "\n",
    "### For a technical audience, usually use a CSV file.\n",
    "\n",
    "If sharing data with other data workers, who are likely to use Python, R, or some similarly nerdy tool, you probably want to create a CSV file.  The reason is simple: You know that this is one of the easiest file types to import in your code, so make life easy for your coworkers, too.  Just call `pd.to_csv( 'my-filename.csv' )` to save your DataFrame.  Although you can use the `sep=\"\\t\"` parameter to create a TSV file, this is rarely what your coworkers want, so it's generally to be avoided.\n",
    "\n",
    "But note that you can lose a lot of important information this way!  You may be familiar with how Excel complains if you try to save an Excel workbook in CSV format, letting you know that you're losing information, such as formatting and formulas.  Any information in your DataFrame other than the text contents of the cells will be lost when saving as CSV.  For instance, if you've converted a column to a categorial variable, that won't be obvious when the data is saved to CSV, and it will be re-imported as plain text.\n",
    "\n",
    "For that reason, we have the following option.\n",
    "\n",
    "### For archiving your own work, use a Pickle file.\n",
    "\n",
    "Python has always had a way to store any Python object in a file, perfectly intact for later loading, using the Pickle format.  The standard extension for this is `.pkl`.  (That's P-K-L, not P-K-one, because it's short for PicKLe.)  The name comes, of course, from the fact that pickling vegetables stores them on the shelf long-term, and yet when you eventually open them later, they're fine.  Similarly, you can store Python objects in a file long-term, open them later, and they're fine.\n",
    "\n",
    "Because Python guarantees that any object you pickle to a file will come back from that file in exactly the same form, you can pickle entire DataFrames and know that every little detail will be preserved, even things that won't get saved correctly to CSV or Excel files, like categorical data types.\n",
    "\n",
    "This is a great way to obey the advice [at the end of the Chapter 11 notes](chapter-11-processing-rows.html#when-the-bottleneck-is-the-dataset).  If you load a big dataset and do a bunch of data cleaning work, and your code is a little slow to run, just save your work to a file right then.\n",
    "\n",
    "```python\n",
    "df.to_pickle( 'cleaned-dataset.pkl' )\n",
    "```\n",
    "\n",
    "Then start a new Python script or Jupyter notebook and load the DataFrame you just saved.\n",
    "\n",
    "```python\n",
    "df = pd.read_pickle( 'cleaned-dataset.pkl' )\n",
    "```\n",
    "\n",
    "Now do all your analysis work in that second script or notebook, and whenever you have to re-run your analysis from the beginning, you won't have to wait for all the data cleaning code to get run again.\n",
    "\n",
    "We won't discuss in these notes the creation of HTML or JSON files from Python.  Although there is occasional value in it, it's much less commonly useful than reading those formats, which we covered above.  We also won't discuss the creation of SQL databases, but here are three ways you can learn more about that.\n",
    " 1. SQL queries can be used to create or modify tables, and we'll see a bit about running SQL queries through Python in the upcoming DataCamp homework.\n",
    " 2. Those interested in learning SQL deeply can take Bentley's CS350 course.\n",
    " 3. Consider forming a team for one of the Learning On Your Own activities shown below."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```{admonition} Learning on Your Own - SQL in Jupyter\n",
    "---\n",
    "class: alert alert-danger\n",
    "---\n",
    "Look up the `ipython-sql` extension to Jupyter and research the following essential parts, then report on them in some sensible medium for your classmates.\n",
    " 1. How to install it and load it.\n",
    " 2. How to connect to a database.\n",
    " 3. How to use both `%sql` and `%%sql` commands.\n",
    " 4. How to store the results of database queries in pandas DataFrames.\n",
    "```\n",
    "\n",
    "```{admonition} Learning on Your Own - SQLite in Python\n",
    "---\n",
    "class: alert alert-danger\n",
    "---\n",
    "Python actually comes with a built-in SQLite database module, which you can use by doing `import sqlite3 as sl`, without even any installation step.  Check out [this blog post](https://towardsdatascience.com/do-you-know-python-has-a-built-in-database-d553989c87bd) for more information, and report to the class on its key features.  Ensure that you include at least one complete example of creating a database on disk and then reading values out of it again later with some queries.\n",
    "```\n",
    "\n",
    "Now that you're familiar with JSON and APIs through your DataCamp prep work for this class, you can also consider some Learning On Your Own opportunities with data for sports analytics.  Here are three examples.\n",
    "\n",
    "```{admonition} Learning on Your Own - College Football Data Python API\n",
    "---\n",
    "class: alert alert-danger\n",
    "---\n",
    "[This tutorial](https://gist.github.com/GJMcClintock/29330e26f674d4103c4f7080212a9208) introduces a Python API for accessing college football data.  Investigate the API and write a report that covers:\n",
    "\n",
    " * How to get the requisite tools.  (E.g., do we have to install or import anything?  How?)\n",
    " * What are the most common functions users will probably want to call, and what kind of data do they yield?\n",
    " * Show a small example investigation to illustrate how the tool could be used, such as creating some visualizations or pivot tables that show something you find interesting in the college football data.\n",
    "```\n",
    "\n",
    "```{admonition} Learning on Your Own - NBA Data Processing Tutorials\n",
    "---\n",
    "class: alert alert-danger\n",
    "---\n",
    "[These tutorials](https://github.com/rd11490/NBA_Tutorials) show a lot of methods for querying basketball data from the web.  But they are very specific to the things that the author found useful for their own work, such as underdog strategies, RAPM, and so on.  Create a report that gives a few smaller, more basic tutorials, answering more introductory questions such as the following (or similar ones you come up with).\n",
    "\n",
    " * What kinds of NBA data exists on the web that we can query with Python tools?\n",
    " * For each of the 3 or 4 most common types of data in your answer to the previous question, do all of the following things:\n",
    "    * Show a short snippet of Python code for fetching that type of data.\n",
    "    * Ensure that you explain that Python code and how the reader might customize it to their own needs.\n",
    "    * Explain the content of the resulting data to whatever degree you are able and/or link to documentation online that provides a data dictionary when possible.\n",
    "    * Do something simple and interesting with the output, such as make a plot or pivot table.\n",
    "```\n",
    "\n",
    "For those interested in sports analytics, [the Python `sportsreference` module](https://pypi.org/project/sportsreference) also seems excellent, but I do not have any LOYOs to assign based on it, because it seems to come with very comprehensive documentation in the form of well-written blog posts.\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "name": "python3",
   "display_name": "Python 3",
   "language": "Python"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}