create_datasets_nyc-tlc.yellow.trips.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1> Explore and create ML datasets </h1>\n",
    "\n",
    "In this notebook, we will explore data corresponding to taxi rides in New York City to build a Machine Learning model in support of a fare-estimation tool. The idea is to suggest a likely fare to taxi riders so that they are not surprised, and so that they can protest if the charge is much higher than expected.\n",
    "\n",
    "<div id=\"toc\"></div>\n",
    "\n",
    "Let's start off with the Python imports that we need."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import google.datalab.bigquery as bq\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import shutil"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3> Extract sample data from BigQuery </h3>\n",
    "\n",
    "The dataset that we will use is <a href=\"https://bigquery.cloud.google.com/table/nyc-tlc:yellow.trips\">a BigQuery public dataset</a>. Click on the link, and look at the column names. Switch to the Details tab to verify that the number of records is one billion, and then switch to the Preview tab to look at a few rows.\n",
    "\n",
    "Write a SQL query to pick up the following fields\n",
    "<pre>\n",
    "  pickup_datetime,\n",
    "  pickup_longitude, pickup_latitude, \n",
    "  dropoff_longitude, dropoff_latitude,\n",
    "  passenger_count,\n",
    "  trip_distance,\n",
    "  tolls_amount,\n",
    "  fare_amount,\n",
    "  total_amount\n",
    "</pre>\n",
    "from the dataset and explore a small part of the data. Make sure to pick a repeatable subset of the data so that if someone reruns this notebook, they will get the same results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "rawdata = \"\"\"\n",
    "SELECT\n",
    "  pickup_datetime,\n",
    "  pickup_longitude, pickup_latitude, \n",
    "  dropoff_longitude, dropoff_latitude,\n",
    "  passenger_count,\n",
    "  trip_distance,\n",
    "  tolls_amount,\n",
    "  fare_amount,\n",
    "  total_amount\n",
    "FROM\n",
    "  `nyc-tlc.yellow.trips`\n",
    "WHERE\n",
    "  MOD(ABS(FARM_FINGERPRINT(CAST(pickup_datetime AS STRING))),EVERY_N) = 1\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "SELECT\n",
      "  pickup_datetime,\n",
      "  pickup_longitude, pickup_latitude, \n",
      "  dropoff_longitude, dropoff_latitude,\n",
      "  passenger_count,\n",
      "  trip_distance,\n",
      "  tolls_amount,\n",
      "  fare_amount,\n",
      "  total_amount\n",
      "FROM\n",
      "  `nyc-tlc.yellow.trips`\n",
      "WHERE\n",
      "  MOD(ABS(FARM_FINGERPRINT(CAST(pickup_datetime AS STRING))),100000) = 1\n",
      "\n",
      "Total dataset is 10789 taxi rides\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>pickup_datetime</th>\n",
       "      <th>pickup_longitude</th>\n",
       "      <th>pickup_latitude</th>\n",
       "      <th>dropoff_longitude</th>\n",
       "      <th>dropoff_latitude</th>\n",
       "      <th>passenger_count</th>\n",
       "      <th>trip_distance</th>\n",
       "      <th>tolls_amount</th>\n",
       "      <th>fare_amount</th>\n",
       "      <th>total_amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2009-11-20 10:54:26</td>\n",
       "      <td>-73.963112</td>\n",
       "      <td>40.758715</td>\n",
       "      <td>-74.004588</td>\n",
       "      <td>40.710867</td>\n",
       "      <td>2</td>\n",
       "      <td>5.80</td>\n",
       "      <td>0.0</td>\n",
       "      <td>15.3</td>\n",
       "      <td>15.80</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2011-12-03 10:28:00</td>\n",
       "      <td>-73.998822</td>\n",
       "      <td>40.680933</td>\n",
       "      <td>-73.968960</td>\n",
       "      <td>40.757878</td>\n",
       "      <td>1</td>\n",
       "      <td>8.28</td>\n",
       "      <td>0.0</td>\n",
       "      <td>20.9</td>\n",
       "      <td>25.58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2012-05-05 22:46:05</td>\n",
       "      <td>-74.009790</td>\n",
       "      <td>40.712483</td>\n",
       "      <td>-73.959293</td>\n",
       "      <td>40.768908</td>\n",
       "      <td>1</td>\n",
       "      <td>5.20</td>\n",
       "      <td>0.0</td>\n",
       "      <td>16.9</td>\n",
       "      <td>19.90</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2010-12-21 13:08:00</td>\n",
       "      <td>-73.982422</td>\n",
       "      <td>40.739847</td>\n",
       "      <td>-73.981658</td>\n",
       "      <td>40.768732</td>\n",
       "      <td>2</td>\n",
       "      <td>2.64</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14.9</td>\n",
       "      <td>15.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2011-12-15 23:27:26</td>\n",
       "      <td>-73.862671</td>\n",
       "      <td>40.768860</td>\n",
       "      <td>-73.983170</td>\n",
       "      <td>40.696372</td>\n",
       "      <td>1</td>\n",
       "      <td>10.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>22.5</td>\n",
       "      <td>27.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2009-03-28 20:30:35</td>\n",
       "      <td>-73.973926</td>\n",
       "      <td>40.757725</td>\n",
       "      <td>-73.981695</td>\n",
       "      <td>40.761591</td>\n",
       "      <td>1</td>\n",
       "      <td>0.50</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2009-08-23 23:59:22</td>\n",
       "      <td>-73.783319</td>\n",
       "      <td>40.648480</td>\n",
       "      <td>-73.893649</td>\n",
       "      <td>40.646566</td>\n",
       "      <td>1</td>\n",
       "      <td>8.30</td>\n",
       "      <td>0.0</td>\n",
       "      <td>18.9</td>\n",
       "      <td>18.90</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>2011-06-04 02:52:10</td>\n",
       "      <td>-73.984681</td>\n",
       "      <td>40.769893</td>\n",
       "      <td>-74.007312</td>\n",
       "      <td>40.705326</td>\n",
       "      <td>1</td>\n",
       "      <td>5.30</td>\n",
       "      <td>0.0</td>\n",
       "      <td>15.3</td>\n",
       "      <td>16.30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>2015-02-17 21:29:49</td>\n",
       "      <td>-73.976730</td>\n",
       "      <td>40.754822</td>\n",
       "      <td>-73.946434</td>\n",
       "      <td>40.725193</td>\n",
       "      <td>2</td>\n",
       "      <td>7.12</td>\n",
       "      <td>0.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>30.36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2014-10-28 06:59:59</td>\n",
       "      <td>-73.953121</td>\n",
       "      <td>40.791453</td>\n",
       "      <td>-73.974214</td>\n",
       "      <td>40.739858</td>\n",
       "      <td>1</td>\n",
       "      <td>4.70</td>\n",
       "      <td>0.0</td>\n",
       "      <td>19.5</td>\n",
       "      <td>25.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      pickup_datetime  pickup_longitude  pickup_latitude  dropoff_longitude  \\\n",
       "0 2009-11-20 10:54:26        -73.963112        40.758715         -74.004588   \n",
       "1 2011-12-03 10:28:00        -73.998822        40.680933         -73.968960   \n",
       "2 2012-05-05 22:46:05        -74.009790        40.712483         -73.959293   \n",
       "3 2010-12-21 13:08:00        -73.982422        40.739847         -73.981658   \n",
       "4 2011-12-15 23:27:26        -73.862671        40.768860         -73.983170   \n",
       "5 2009-03-28 20:30:35        -73.973926        40.757725         -73.981695   \n",
       "6 2009-08-23 23:59:22        -73.783319        40.648480         -73.893649   \n",
       "7 2011-06-04 02:52:10        -73.984681        40.769893         -74.007312   \n",
       "8 2015-02-17 21:29:49        -73.976730        40.754822         -73.946434   \n",
       "9 2014-10-28 06:59:59        -73.953121        40.791453         -73.974214   \n",
       "\n",
       "   dropoff_latitude  passenger_count  trip_distance  tolls_amount  \\\n",
       "0         40.710867                2           5.80           0.0   \n",
       "1         40.757878                1           8.28           0.0   \n",
       "2         40.768908                1           5.20           0.0   \n",
       "3         40.768732                2           2.64           0.0   \n",
       "4         40.696372                1          10.00           0.0   \n",
       "5         40.761591                1           0.50           0.0   \n",
       "6         40.646566                1           8.30           0.0   \n",
       "7         40.705326                1           5.30           0.0   \n",
       "8         40.725193                2           7.12           0.0   \n",
       "9         40.739858                1           4.70           0.0   \n",
       "\n",
       "   fare_amount  total_amount  \n",
       "0         15.3         15.80  \n",
       "1         20.9         25.58  \n",
       "2         16.9         19.90  \n",
       "3         14.9         15.40  \n",
       "4         22.5         27.00  \n",
       "5          4.6          4.60  \n",
       "6         18.9         18.90  \n",
       "7         15.3         16.30  \n",
       "8         24.0         30.36  \n",
       "9         19.5         25.00  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = rawdata.replace(\"EVERY_N\", \"100000\")\n",
    "print(query)\n",
    "trips = bq.Query(query).execute().result().to_dataframe()\n",
    "print(\"Total dataset is {} taxi rides\".format(len(trips)))\n",
    "trips[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3> Exploring data </h3>\n",
    "\n",
    "Let's explore this dataset and clean it up as necessary. We'll use the Python Seaborn package to visualize graphs and Pandas to do the slicing and filtering."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/envs/py3env/lib/python3.5/site-packages/matplotlib/font_manager.py:1320: UserWarning: findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans\n",
      "  (prop.get_family(), self.defaultFamily[fontext]))\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f50cc1c27f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = sns.regplot(x = \"trip_distance\", y = \"fare_amount\", ci = None, truncate = True, data = trips)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Hmm ... do you see something wrong with the data that needs addressing?\n",
    "\n",
    "It appears that we have a lot of invalid data that is being coded as zero distance and some fare amounts that are definitely illegitimate. Let's remove them from our analysis. We can do this by modifying the BigQuery query to keep only trips longer than zero miles and fare amounts that are at least the minimum cab fare ($2.50).\n",
    "\n",
    "What's up with the streaks at \\$45 and \\$50?  Those are fixed-amount rides from JFK and La Guardia airports into anywhere in Manhattan, i.e. to be expected. Let's list the data to make sure the values look reasonable.\n",
    "\n",
    "Let's examine whether the toll amount is captured in the total amount."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>pickup_datetime</th>\n",
       "      <th>pickup_longitude</th>\n",
       "      <th>pickup_latitude</th>\n",
       "      <th>dropoff_longitude</th>\n",
       "      <th>dropoff_latitude</th>\n",
       "      <th>passenger_count</th>\n",
       "      <th>trip_distance</th>\n",
       "      <th>tolls_amount</th>\n",
       "      <th>fare_amount</th>\n",
       "      <th>total_amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-73.873527</td>\n",
       "      <td>40.773922</td>\n",
       "      <td>-73.944947</td>\n",
       "      <td>40.828902</td>\n",
       "      <td>5</td>\n",
       "      <td>8.35</td>\n",
       "      <td>5.33</td>\n",
       "      <td>25.0</td>\n",
       "      <td>36.33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>153</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-73.874567</td>\n",
       "      <td>40.774102</td>\n",
       "      <td>-73.986480</td>\n",
       "      <td>40.759202</td>\n",
       "      <td>1</td>\n",
       "      <td>11.43</td>\n",
       "      <td>5.33</td>\n",
       "      <td>37.0</td>\n",
       "      <td>54.04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>242</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>6.38</td>\n",
       "      <td>5.33</td>\n",
       "      <td>20.0</td>\n",
       "      <td>26.33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>325</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-73.781987</td>\n",
       "      <td>40.644755</td>\n",
       "      <td>-73.921672</td>\n",
       "      <td>40.828412</td>\n",
       "      <td>1</td>\n",
       "      <td>19.08</td>\n",
       "      <td>5.33</td>\n",
       "      <td>50.5</td>\n",
       "      <td>56.83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>558</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-74.005625</td>\n",
       "      <td>40.736770</td>\n",
       "      <td>-73.919840</td>\n",
       "      <td>40.877687</td>\n",
       "      <td>1</td>\n",
       "      <td>14.33</td>\n",
       "      <td>2.44</td>\n",
       "      <td>42.0</td>\n",
       "      <td>54.43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>678</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-73.870160</td>\n",
       "      <td>40.760742</td>\n",
       "      <td>-73.980912</td>\n",
       "      <td>40.788492</td>\n",
       "      <td>4</td>\n",
       "      <td>10.37</td>\n",
       "      <td>5.33</td>\n",
       "      <td>30.5</td>\n",
       "      <td>45.91</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1156</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-73.984350</td>\n",
       "      <td>40.761760</td>\n",
       "      <td>-73.859030</td>\n",
       "      <td>40.718510</td>\n",
       "      <td>2</td>\n",
       "      <td>10.04</td>\n",
       "      <td>5.33</td>\n",
       "      <td>31.5</td>\n",
       "      <td>37.83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1476</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-73.873210</td>\n",
       "      <td>40.774042</td>\n",
       "      <td>-73.978013</td>\n",
       "      <td>40.755800</td>\n",
       "      <td>5</td>\n",
       "      <td>10.45</td>\n",
       "      <td>5.33</td>\n",
       "      <td>30.5</td>\n",
       "      <td>36.83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1527</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-73.874445</td>\n",
       "      <td>40.774037</td>\n",
       "      <td>-73.982382</td>\n",
       "      <td>40.784930</td>\n",
       "      <td>1</td>\n",
       "      <td>9.97</td>\n",
       "      <td>5.33</td>\n",
       "      <td>30.5</td>\n",
       "      <td>44.10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4190</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-73.975610</td>\n",
       "      <td>40.751660</td>\n",
       "      <td>-73.953920</td>\n",
       "      <td>40.743240</td>\n",
       "      <td>1</td>\n",
       "      <td>2.22</td>\n",
       "      <td>5.33</td>\n",
       "      <td>8.5</td>\n",
       "      <td>14.83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5532</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-73.987960</td>\n",
       "      <td>40.754062</td>\n",
       "      <td>-73.956627</td>\n",
       "      <td>40.745962</td>\n",
       "      <td>2</td>\n",
       "      <td>3.11</td>\n",
       "      <td>5.33</td>\n",
       "      <td>11.5</td>\n",
       "      <td>21.30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6612</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-73.783898</td>\n",
       "      <td>40.648660</td>\n",
       "      <td>-73.981845</td>\n",
       "      <td>40.762872</td>\n",
       "      <td>6</td>\n",
       "      <td>18.18</td>\n",
       "      <td>5.33</td>\n",
       "      <td>52.0</td>\n",
       "      <td>69.30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6811</th>\n",
       "      <td>2014-05-20 23:09:00</td>\n",
       "      <td>-74.011087</td>\n",
       "      <td>40.724797</td>\n",
       "      <td>-74.026347</td>\n",
       "      <td>40.750997</td>\n",
       "      <td>1</td>\n",
       "      <td>4.09</td>\n",
       "      <td>11.00</td>\n",
       "      <td>45.0</td>\n",
       "      <td>61.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         pickup_datetime  pickup_longitude  pickup_latitude  \\\n",
       "92   2014-05-20 23:09:00        -73.873527        40.773922   \n",
       "153  2014-05-20 23:09:00        -73.874567        40.774102   \n",
       "242  2014-05-20 23:09:00          0.000000         0.000000   \n",
       "325  2014-05-20 23:09:00        -73.781987        40.644755   \n",
       "558  2014-05-20 23:09:00        -74.005625        40.736770   \n",
       "678  2014-05-20 23:09:00        -73.870160        40.760742   \n",
       "1156 2014-05-20 23:09:00        -73.984350        40.761760   \n",
       "1476 2014-05-20 23:09:00        -73.873210        40.774042   \n",
       "1527 2014-05-20 23:09:00        -73.874445        40.774037   \n",
       "4190 2014-05-20 23:09:00        -73.975610        40.751660   \n",
       "5532 2014-05-20 23:09:00        -73.987960        40.754062   \n",
       "6612 2014-05-20 23:09:00        -73.783898        40.648660   \n",
       "6811 2014-05-20 23:09:00        -74.011087        40.724797   \n",
       "\n",
       "      dropoff_longitude  dropoff_latitude  passenger_count  trip_distance  \\\n",
       "92           -73.944947         40.828902                5           8.35   \n",
       "153          -73.986480         40.759202                1          11.43   \n",
       "242            0.000000          0.000000                1           6.38   \n",
       "325          -73.921672         40.828412                1          19.08   \n",
       "558          -73.919840         40.877687                1          14.33   \n",
       "678          -73.980912         40.788492                4          10.37   \n",
       "1156         -73.859030         40.718510                2          10.04   \n",
       "1476         -73.978013         40.755800                5          10.45   \n",
       "1527         -73.982382         40.784930                1           9.97   \n",
       "4190         -73.953920         40.743240                1           2.22   \n",
       "5532         -73.956627         40.745962                2           3.11   \n",
       "6612         -73.981845         40.762872                6          18.18   \n",
       "6811         -74.026347         40.750997                1           4.09   \n",
       "\n",
       "      tolls_amount  fare_amount  total_amount  \n",
       "92            5.33         25.0         36.33  \n",
       "153           5.33         37.0         54.04  \n",
       "242           5.33         20.0         26.33  \n",
       "325           5.33         50.5         56.83  \n",
       "558           2.44         42.0         54.43  \n",
       "678           5.33         30.5         45.91  \n",
       "1156          5.33         31.5         37.83  \n",
       "1476          5.33         30.5         36.83  \n",
       "1527          5.33         30.5         44.10  \n",
       "4190          5.33          8.5         14.83  \n",
       "5532          5.33         11.5         21.30  \n",
       "6612          5.33         52.0         69.30  \n",
       "6811         11.00         45.0         61.00  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tollrides = trips[trips['tolls_amount'] > 0]\n",
    "tollrides[tollrides['pickup_datetime'] == '2014-05-20 23:09:00']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looking a few samples above, it should be clear that the total amount reflects fare amount, toll and tip somewhat arbitrarily -- this is because when customers pay cash, the tip is not known.  So, we'll use the sum of fare_amount + tolls_amount as what needs to be predicted.  Tips are discretionary and do not have to be included in our fare estimation tool.\n",
    "\n",
    "Let's also look at the distribution of values within the columns."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>pickup_longitude</th>\n",
       "      <th>pickup_latitude</th>\n",
       "      <th>dropoff_longitude</th>\n",
       "      <th>dropoff_latitude</th>\n",
       "      <th>passenger_count</th>\n",
       "      <th>trip_distance</th>\n",
       "      <th>tolls_amount</th>\n",
       "      <th>fare_amount</th>\n",
       "      <th>total_amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>10789.000000</td>\n",
       "      <td>10789.000000</td>\n",
       "      <td>10789.000000</td>\n",
       "      <td>10789.000000</td>\n",
       "      <td>10789.000000</td>\n",
       "      <td>10789.000000</td>\n",
       "      <td>10789.000000</td>\n",
       "      <td>10789.000000</td>\n",
       "      <td>10789.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>-72.577466</td>\n",
       "      <td>39.988518</td>\n",
       "      <td>-72.577056</td>\n",
       "      <td>39.991694</td>\n",
       "      <td>1.648716</td>\n",
       "      <td>2.837068</td>\n",
       "      <td>0.228359</td>\n",
       "      <td>11.165192</td>\n",
       "      <td>13.280564</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>10.071200</td>\n",
       "      <td>5.523964</td>\n",
       "      <td>10.068592</td>\n",
       "      <td>5.510504</td>\n",
       "      <td>1.282589</td>\n",
       "      <td>3.319036</td>\n",
       "      <td>1.144977</td>\n",
       "      <td>9.444876</td>\n",
       "      <td>11.267704</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-75.336579</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-75.381754</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.500000</td>\n",
       "      <td>2.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>-73.992172</td>\n",
       "      <td>40.735837</td>\n",
       "      <td>-73.991581</td>\n",
       "      <td>40.734267</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.010000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>7.200000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>-73.981871</td>\n",
       "      <td>40.753200</td>\n",
       "      <td>-73.980401</td>\n",
       "      <td>40.752872</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.750000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>8.500000</td>\n",
       "      <td>10.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>-73.967330</td>\n",
       "      <td>40.767309</td>\n",
       "      <td>-73.964102</td>\n",
       "      <td>40.767484</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>3.110000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>12.500000</td>\n",
       "      <td>14.600000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>41.366138</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>41.366138</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>42.800000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>194.750000</td>\n",
       "      <td>204.750000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       pickup_longitude  pickup_latitude  dropoff_longitude  dropoff_latitude  \\\n",
       "count      10789.000000     10789.000000       10789.000000      10789.000000   \n",
       "mean         -72.577466        39.988518         -72.577056         39.991694   \n",
       "std           10.071200         5.523964          10.068592          5.510504   \n",
       "min          -75.336579         0.000000         -75.381754          0.000000   \n",
       "25%          -73.992172        40.735837         -73.991581         40.734267   \n",
       "50%          -73.981871        40.753200         -73.980401         40.752872   \n",
       "75%          -73.967330        40.767309         -73.964102         40.767484   \n",
       "max            0.000000        41.366138           0.000000         41.366138   \n",
       "\n",
       "       passenger_count  trip_distance  tolls_amount   fare_amount  \\\n",
       "count     10789.000000   10789.000000  10789.000000  10789.000000   \n",
       "mean          1.648716       2.837068      0.228359     11.165192   \n",
       "std           1.282589       3.319036      1.144977      9.444876   \n",
       "min           0.000000       0.000000      0.000000      2.500000   \n",
       "25%           1.000000       1.010000      0.000000      6.000000   \n",
       "50%           1.000000       1.750000      0.000000      8.500000   \n",
       "75%           2.000000       3.110000      0.000000     12.500000   \n",
       "max           6.000000      42.800000     16.000000    194.750000   \n",
       "\n",
       "       total_amount  \n",
       "count  10789.000000  \n",
       "mean      13.280564  \n",
       "std       11.267704  \n",
       "min        2.500000  \n",
       "25%        7.200000  \n",
       "50%       10.000000  \n",
       "75%       14.600000  \n",
       "max      204.750000  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trips.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Hmm ... The min, max of longitude look strange.\n",
    "\n",
    "Finally, let's actually look at the start and end of a few of the trips."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/envs/py3env/lib/python3.5/site-packages/matplotlib/font_manager.py:1320: UserWarning: findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans\n",
      "  (prop.get_family(), self.defaultFamily[fontext]))\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f50cbb939e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def showrides(df, numlines):\n",
    "  import matplotlib.pyplot as plt\n",
    "  lats = []\n",
    "  lons = []\n",
    "  goodrows = df[df['pickup_longitude'] < -70]\n",
    "  for iter, row in goodrows[:numlines].iterrows():\n",
    "    lons.append(row['pickup_longitude'])\n",
    "    lons.append(row['dropoff_longitude'])\n",
    "    lons.append(None)\n",
    "    lats.append(row['pickup_latitude'])\n",
    "    lats.append(row['dropoff_latitude'])\n",
    "    lats.append(None)\n",
    "\n",
    "  sns.set_style(\"darkgrid\")\n",
    "  plt.plot(lons, lats)\n",
    "\n",
    "showrides(trips, 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/envs/py3env/lib/python3.5/site-packages/matplotlib/font_manager.py:1320: UserWarning: findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans\n",
      "  (prop.get_family(), self.defaultFamily[fontext]))\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f50cbba6518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "showrides(tollrides, 10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you'd expect, rides that involve a toll are longer than the typical ride."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3> Quality control and other preprocessing </h3>\n",
    "\n",
    "We need to do some clean-up of the data:\n",
    "<ol>\n",
    "<li>New York city longitudes are around -74 and latitudes are around 41.</li>\n",
    "<li>We shouldn't have zero passengers.</li>\n",
    "<li>Clean up the total_amount column to reflect only fare_amount and tolls_amount, and then remove those two columns.</li>\n",
    "<li>Before the ride starts, we'll know the pickup and dropoff locations, but not the trip distance (that depends on the route taken), so remove it from the ML dataset</li>\n",
    "<li>Discard the timestamp</li>\n",
    "</ol>\n",
    "\n",
    "Let's change the BigQuery query appropriately.  In production, we'll have to carry out the same preprocessing on the real-time input data. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "SELECT\n",
      "  (tolls_amount + fare_amount) AS fare_amount,\n",
      "  pickup_longitude AS pickuplon,\n",
      "  pickup_latitude AS pickuplat,\n",
      "  dropoff_longitude AS dropofflon,\n",
      "  dropoff_latitude AS dropofflat,\n",
      "  passenger_count*1.0 AS passengers\n",
      "FROM\n",
      "  `nyc-tlc.yellow.trips`\n",
      "WHERE\n",
      "  trip_distance > 0\n",
      "  AND fare_amount >= 2.5\n",
      "  AND pickup_longitude > -78\n",
      "  AND pickup_longitude < -70\n",
      "  AND dropoff_longitude > -78\n",
      "  AND dropoff_longitude < -70\n",
      "  AND pickup_latitude > 37\n",
      "  AND pickup_latitude < 45\n",
      "  AND dropoff_latitude > 37\n",
      "  AND dropoff_latitude < 45\n",
      "  AND passenger_count > 0\n",
      "  \n",
      "AND MOD(ABS(FARM_FINGERPRINT(CAST(pickup_datetime AS STRING))), 100000) = 1\n",
      "AND MOD(ABS(FARM_FINGERPRINT(CAST(pickup_datetime AS STRING))), 100000 * 100) >= (100000 * 0)\n",
      " AND MOD(ABS(FARM_FINGERPRINT(CAST(pickup_datetime AS STRING))), 100000 * 100) < (100000 * 70)\n"
     ]
    }
   ],
   "source": [
    "def sample_between(a, b):\n",
    "    basequery = \"\"\"\n",
    "SELECT\n",
    "  (tolls_amount + fare_amount) AS fare_amount,\n",
    "  pickup_longitude AS pickuplon,\n",
    "  pickup_latitude AS pickuplat,\n",
    "  dropoff_longitude AS dropofflon,\n",
    "  dropoff_latitude AS dropofflat,\n",
    "  passenger_count*1.0 AS passengers\n",
    "FROM\n",
    "  `nyc-tlc.yellow.trips`\n",
    "WHERE\n",
    "  trip_distance > 0\n",
    "  AND fare_amount >= 2.5\n",
    "  AND pickup_longitude > -78\n",
    "  AND pickup_longitude < -70\n",
    "  AND dropoff_longitude > -78\n",
    "  AND dropoff_longitude < -70\n",
    "  AND pickup_latitude > 37\n",
    "  AND pickup_latitude < 45\n",
    "  AND dropoff_latitude > 37\n",
    "  AND dropoff_latitude < 45\n",
    "  AND passenger_count > 0\n",
    "  \"\"\"\n",
    "    sampler = \"AND MOD(ABS(FARM_FINGERPRINT(CAST(pickup_datetime AS STRING))), EVERY_N) = 1\"\n",
    "    sampler2 = \"AND {0} >= {1}\\n AND {0} < {2}\".format(\n",
    "           \"MOD(ABS(FARM_FINGERPRINT(CAST(pickup_datetime AS STRING))), EVERY_N * 100)\",\n",
    "           \"(EVERY_N * {})\".format(a), \"(EVERY_N * {})\".format(b)\n",
    "        )\n",
    "    return \"{}\\n{}\\n{}\".format(basequery, sampler, sampler2)\n",
    "\n",
    "def create_query(phase, EVERY_N):\n",
    "  \"\"\"Phase: train (70%) valid (15%) or test (15%)\"\"\"\n",
    "  query = \"\"\n",
    "  if phase == 'train':\n",
    "    # Training\n",
    "    query = sample_between(0, 70)\n",
    "  elif phase == 'valid':\n",
    "    # Validation\n",
    "    query = sample_between(70, 85)\n",
    "  else:\n",
    "    # Test\n",
    "    query = sample_between(85, 100)\n",
    "  return query.replace(\"EVERY_N\", str(EVERY_N))\n",
    "\n",
    "print(create_query('train', 100000))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      fare_amount  pickuplon  pickuplat  dropofflon  dropofflat  passengers\n",
      "0            4.90 -73.986226  40.747450  -73.997979   40.750835         1.0\n",
      "1            5.30 -73.984751  40.732240  -73.977915   40.744018         1.0\n",
      "2            8.50 -74.000473  40.730095  -73.983398   40.722160         1.0\n",
      "3            5.30 -73.972066  40.745835  -73.971302   40.760418         1.0\n",
      "4           12.10 -73.993161  40.692794  -74.010602   40.728559         1.0\n",
      "5            6.50 -73.881155  40.756119  -73.890311   40.747039         1.0\n",
      "6            8.00 -74.012676  40.701672  -74.016216   40.716223         2.0\n",
      "7           13.50 -73.973197  40.755287  -73.990513   40.761235         4.0\n",
      "8            3.30 -73.950739  40.779368  -73.955884   40.779298         1.0\n",
      "9            6.10 -73.974348  40.753430  -73.990849   40.761080         1.0\n",
      "10           6.90 -73.992268  40.749388  -73.983942   40.754445         1.0\n",
      "11          16.90 -74.009790  40.712483  -73.959293   40.768908         1.0\n",
      "12           7.50 -73.989090  40.748477  -73.982650   40.761237         1.0\n",
      "13           6.50 -74.011482  40.720081  -73.999519   40.722826         1.0\n",
      "14           8.50 -73.986830  40.746070  -74.006891   40.716072         1.0\n",
      "15          13.70 -73.988733  40.777815  -74.009841   40.721087         1.0\n",
      "16          24.00 -73.976730  40.754822  -73.946434   40.725193         2.0\n",
      "17           9.00 -73.992091  40.769537  -73.976841   40.761546         1.0\n",
      "18          18.90 -73.783319  40.648480  -73.893649   40.646566         1.0\n",
      "19           6.10 -73.985611  40.741404  -74.002782   40.738735         1.0\n",
      "20           8.50 -73.978104  40.745924  -73.985299   40.761861         1.0\n",
      "21          10.90 -73.991950  40.726075  -73.961097   40.768953         1.0\n",
      "22           7.70 -73.994739  40.725984  -73.970818   40.755763         1.0\n",
      "23           9.50 -73.978853  40.760868  -73.985674   40.749288         1.0\n",
      "24          12.10 -73.990708  40.770127  -73.989685   40.734277         1.0\n",
      "25          22.50 -73.862671  40.768860  -73.983170   40.696372         1.0\n",
      "26           7.30 -73.994339  40.766466  -73.996446   40.753076         2.0\n",
      "27           8.90 -73.997208  40.762527  -73.996209   40.737100         2.0\n",
      "28          14.90 -73.980313  40.744282  -74.015285   40.711458         1.0\n",
      "29          14.10 -73.961667  40.776937  -73.999829   40.742608         1.0\n",
      "...           ...        ...        ...         ...         ...         ...\n",
      "987          3.70 -73.982385  40.731488  -73.976898   40.738970         1.0\n",
      "988         12.50 -73.949934  40.780572  -73.983804   40.748626         1.0\n",
      "989         15.50 -73.972139  40.756511  -73.983312   40.773414         1.0\n",
      "990          9.30 -73.980882  40.778346  -73.957141   40.781745         1.0\n",
      "991          4.10 -73.982100  40.752211  -73.986818   40.748794         1.0\n",
      "992         10.50 -74.009268  40.713043  -73.994740   40.760764         1.0\n",
      "993          4.90 -73.957898  40.776407  -73.953610   40.790757         1.0\n",
      "994         16.00 -73.989302  40.730052  -73.949112   40.780188         1.0\n",
      "995          8.10 -73.991938  40.753912  -73.981894   40.772891         2.0\n",
      "996          6.90 -73.990019  40.756730  -73.970250   40.756209         1.0\n",
      "997         18.00 -73.871222  40.773470  -73.942462   40.748538         1.0\n",
      "998          9.00 -74.004738  40.741760  -73.985161   40.732368         1.0\n",
      "999         10.10 -73.993957  40.750054  -73.989109   40.736142         3.0\n",
      "1000         6.10 -73.982069  40.745810  -73.988017   40.749831         1.0\n",
      "1001         6.50 -73.973133  40.747505  -73.983042   40.760332         1.0\n",
      "1002         7.80 -73.990772  40.745941  -73.979892   40.773873         1.0\n",
      "1003         8.50 -73.969371  40.760608  -73.958659   40.782875         1.0\n",
      "1004         7.30 -73.979871  40.769958  -73.971427   40.750727         2.0\n",
      "1005         5.00 -74.004195  40.752131  -73.993223   40.748832         1.0\n",
      "1006         8.20 -73.915341  40.743357  -73.915341   40.743357         1.0\n",
      "1007         5.00 -73.989468  40.741820  -73.999620   40.735915         1.0\n",
      "1008        33.33 -73.964058  40.764322  -73.870316   40.773538         1.0\n",
      "1009         4.90 -73.998998  40.761256  -73.995058   40.752831         1.0\n",
      "1010         8.00 -73.999992  40.721725  -74.010086   40.714523         2.0\n",
      "1011        12.00 -73.984623  40.764327  -74.009840   40.721709         1.0\n",
      "1012         5.70 -73.982005  40.755126  -73.984671   40.761793         1.0\n",
      "1013         6.50 -73.970156  40.759188  -73.973848   40.743818         1.0\n",
      "1014         7.00 -74.002987  40.733210  -73.994094   40.739433         1.0\n",
      "1015        19.70 -73.978270  40.788167  -73.992229   40.725066         1.0\n",
      "1016        16.50 -74.010971  40.703091  -73.971237   40.673207         1.0\n",
      "\n",
      "[1017 rows x 6 columns]\n",
      "139984955151192\n",
      "139984955153656\n",
      "['fare_amount', 'pickuplon', 'pickuplat', 'dropofflon', 'dropofflat', 'passengers']\n",
      "['pickuplon', 'pickuplat', 'dropofflon', 'dropofflat', 'passengers']\n",
      "['fare_amount', 'pickuplon', 'pickuplat', 'dropofflon', 'dropofflat', 'passengers']\n",
      "      fare_amount  pickuplon  pickuplat  dropofflon  dropofflat  passengers\n",
      "0            4.90 -73.986226  40.747450  -73.997979   40.750835         1.0\n",
      "1            5.30 -73.984751  40.732240  -73.977915   40.744018         1.0\n",
      "2            8.50 -74.000473  40.730095  -73.983398   40.722160         1.0\n",
      "3            5.30 -73.972066  40.745835  -73.971302   40.760418         1.0\n",
      "4           12.10 -73.993161  40.692794  -74.010602   40.728559         1.0\n",
      "5            6.50 -73.881155  40.756119  -73.890311   40.747039         1.0\n",
      "6            8.00 -74.012676  40.701672  -74.016216   40.716223         2.0\n",
      "7           13.50 -73.973197  40.755287  -73.990513   40.761235         4.0\n",
      "8            3.30 -73.950739  40.779368  -73.955884   40.779298         1.0\n",
      "9            6.10 -73.974348  40.753430  -73.990849   40.761080         1.0\n",
      "10           6.90 -73.992268  40.749388  -73.983942   40.754445         1.0\n",
      "11          16.90 -74.009790  40.712483  -73.959293   40.768908         1.0\n",
      "12           7.50 -73.989090  40.748477  -73.982650   40.761237         1.0\n",
      "13           6.50 -74.011482  40.720081  -73.999519   40.722826         1.0\n",
      "14           8.50 -73.986830  40.746070  -74.006891   40.716072         1.0\n",
      "15          13.70 -73.988733  40.777815  -74.009841   40.721087         1.0\n",
      "16          24.00 -73.976730  40.754822  -73.946434   40.725193         2.0\n",
      "17           9.00 -73.992091  40.769537  -73.976841   40.761546         1.0\n",
      "18          18.90 -73.783319  40.648480  -73.893649   40.646566         1.0\n",
      "19           6.10 -73.985611  40.741404  -74.002782   40.738735         1.0\n",
      "20           8.50 -73.978104  40.745924  -73.985299   40.761861         1.0\n",
      "21          10.90 -73.991950  40.726075  -73.961097   40.768953         1.0\n",
      "22           7.70 -73.994739  40.725984  -73.970818   40.755763         1.0\n",
      "23           9.50 -73.978853  40.760868  -73.985674   40.749288         1.0\n",
      "24          12.10 -73.990708  40.770127  -73.989685   40.734277         1.0\n",
      "25          22.50 -73.862671  40.768860  -73.983170   40.696372         1.0\n",
      "26           7.30 -73.994339  40.766466  -73.996446   40.753076         2.0\n",
      "27           8.90 -73.997208  40.762527  -73.996209   40.737100         2.0\n",
      "28          14.90 -73.980313  40.744282  -74.015285   40.711458         1.0\n",
      "29          14.10 -73.961667  40.776937  -73.999829   40.742608         1.0\n",
      "...           ...        ...        ...         ...         ...         ...\n",
      "987          3.70 -73.982385  40.731488  -73.976898   40.738970         1.0\n",
      "988         12.50 -73.949934  40.780572  -73.983804   40.748626         1.0\n",
      "989         15.50 -73.972139  40.756511  -73.983312   40.773414         1.0\n",
      "990          9.30 -73.980882  40.778346  -73.957141   40.781745         1.0\n",
      "991          4.10 -73.982100  40.752211  -73.986818   40.748794         1.0\n",
      "992         10.50 -74.009268  40.713043  -73.994740   40.760764         1.0\n",
      "993          4.90 -73.957898  40.776407  -73.953610   40.790757         1.0\n",
      "994         16.00 -73.989302  40.730052  -73.949112   40.780188         1.0\n",
      "995          8.10 -73.991938  40.753912  -73.981894   40.772891         2.0\n",
      "996          6.90 -73.990019  40.756730  -73.970250   40.756209         1.0\n",
      "997         18.00 -73.871222  40.773470  -73.942462   40.748538         1.0\n",
      "998          9.00 -74.004738  40.741760  -73.985161   40.732368         1.0\n",
      "999         10.10 -73.993957  40.750054  -73.989109   40.736142         3.0\n",
      "1000         6.10 -73.982069  40.745810  -73.988017   40.749831         1.0\n",
      "1001         6.50 -73.973133  40.747505  -73.983042   40.760332         1.0\n",
      "1002         7.80 -73.990772  40.745941  -73.979892   40.773873         1.0\n",
      "1003         8.50 -73.969371  40.760608  -73.958659   40.782875         1.0\n",
      "1004         7.30 -73.979871  40.769958  -73.971427   40.750727         2.0\n",
      "1005         5.00 -74.004195  40.752131  -73.993223   40.748832         1.0\n",
      "1006         8.20 -73.915341  40.743357  -73.915341   40.743357         1.0\n",
      "1007         5.00 -73.989468  40.741820  -73.999620   40.735915         1.0\n",
      "1008        33.33 -73.964058  40.764322  -73.870316   40.773538         1.0\n",
      "1009         4.90 -73.998998  40.761256  -73.995058   40.752831         1.0\n",
      "1010         8.00 -73.999992  40.721725  -74.010086   40.714523         2.0\n",
      "1011        12.00 -73.984623  40.764327  -74.009840   40.721709         1.0\n",
      "1012         5.70 -73.982005  40.755126  -73.984671   40.761793         1.0\n",
      "1013         6.50 -73.970156  40.759188  -73.973848   40.743818         1.0\n",
      "1014         7.00 -74.002987  40.733210  -73.994094   40.739433         1.0\n",
      "1015        19.70 -73.978270  40.788167  -73.992229   40.725066         1.0\n",
      "1016        16.50 -74.010971  40.703091  -73.971237   40.673207         1.0\n",
      "\n",
      "[1017 rows x 6 columns]\n"
     ]
    }
   ],
   "source": [
    "query = create_query(phase, 100000)\n",
    "df = bq.Query(query).execute().result().to_dataframe()\n",
    "outdf = df.copy(deep = False)\n",
    "print(outdf)\n",
    "print(id(df))\n",
    "print(id(outdf))\n",
    "# Reorder columns so that target is first column\n",
    "cols = outdf.columns.tolist()\n",
    "print(cols)\n",
    "cols.remove('fare_amount')\n",
    "print(cols)\n",
    "cols.insert(0, 'fare_amount')\n",
    "print(cols)\n",
    "print(outdf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['fare_amount', 'pickuplon', 'pickuplat', 'dropofflon', 'dropofflat', 'passengers', 'key']\n",
      "Wrote 7645 to taxi-train.csv\n",
      "['fare_amount', 'pickuplon', 'pickuplat', 'dropofflon', 'dropofflat', 'passengers', 'key']\n",
      "Wrote 1814 to taxi-valid.csv\n",
      "['fare_amount', 'pickuplon', 'pickuplat', 'dropofflon', 'dropofflat', 'passengers', 'key']\n",
      "Wrote 1017 to taxi-test.csv\n"
     ]
    }
   ],
   "source": [
    "def to_csv(df, filename):\n",
    "  outdf = df.copy(deep = False)\n",
    "  outdf.loc[:, 'key'] = np.arange(0, len(outdf)) # rownumber as key\n",
    "  # Reorder columns so that target is first column\n",
    "  cols = outdf.columns.tolist()\n",
    "  cols.remove('fare_amount')\n",
    "  cols.insert(0, 'fare_amount')\n",
    "  print(cols)  # new order of columns\n",
    "  outdf = outdf[cols]\n",
    "  outdf.to_csv(filename, header = False, index_label = False, index = False)\n",
    "  print(\"Wrote {} to {}\".format(len(outdf), filename))\n",
    "\n",
    "for phase in ['train', 'valid', 'test']:\n",
    "  query = create_query(phase, 100000)\n",
    "  df = bq.Query(query).execute().result().to_dataframe()\n",
    "  to_csv(df, 'taxi-{}.csv'.format(phase))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3> Verify that datasets exist </h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-rw-r--r-- 1 root root  58048 May 12 15:04 taxi-test.csv\r\n",
      "-rw-r--r-- 1 root root 434648 May 12 15:03 taxi-train.csv\r\n",
      "-rw-r--r-- 1 root root 102591 May 12 15:03 taxi-valid.csv\r\n"
     ]
    }
   ],
   "source": [
    "!ls -l *.csv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We have 3 .csv files corresponding to train, valid, test.  The ratio of file-sizes correspond to our split of the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6.5,-73.981565,40.732902,-73.989255,40.748087,1.0,0\n",
      "7.5,-73.977417,40.757775,-73.98744,40.745502,1.0,1\n",
      "8.5,-73.952977,40.780286,-73.947307,40.811909,1.0,2\n",
      "9.0,-73.978015,40.76337,-73.981373,40.784517,1.0,3\n",
      "17.0,-73.976362,40.751972,-74.009268,40.701627,4.0,4\n",
      "14.1,-73.98893,40.726817,-73.954809,40.725432,2.0,5\n",
      "31.0,-73.99955,40.7606,-73.99965,40.72522,1.0,6\n",
      "8.1,-73.987746,40.722759,-74.000635,40.740735,2.0,7\n",
      "4.9,-74.01016,40.720395,-74.016025,40.711103,5.0,8\n",
      "11.3,-73.981742,40.775885,-73.955807,40.822712,1.0,9\n"
     ]
    }
   ],
   "source": [
    "%bash\n",
    "head taxi-train.csv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looks good! We now have our ML datasets and are ready to train ML models, validate them and evaluate them."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3> Benchmark </h3>\n",
    "\n",
    "Before we start building complex ML models, it is a good idea to come up with a very simple model and use that as a benchmark.\n",
    "\n",
    "My model is going to be to simply divide the mean fare_amount by the mean trip_distance to come up with a rate and use that to predict.  Let's compute the RMSE of such a model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rate = $2.6164717631130308/km\n",
      "Train RMSE = 7.459449914893905\n",
      "Valid RMSE = 9.354082608239274\n",
      "Test RMSE = 5.444708079525852\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/envs/py3env/lib/python3.5/site-packages/ipykernel/__main__.py:7: RuntimeWarning: invalid value encountered in arccos\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import shutil\n",
    "\n",
    "def distance_between(lat1, lon1, lat2, lon2):\n",
    "  # Haversine formula to compute distance \"as the crow flies\".  Taxis can't fly of course.\n",
    "  dist = np.degrees(np.arccos(np.sin(np.radians(lat1)) * np.sin(np.radians(lat2)) + np.cos(np.radians(lat1)) * np.cos(np.radians(lat2)) * np.cos(np.radians(lon2 - lon1)))) * 60 * 1.515 * 1.609344\n",
    "  return dist\n",
    "\n",
    "def estimate_distance(df):\n",
    "  return distance_between(df['pickuplat'], df['pickuplon'], df['dropofflat'], df['dropofflon'])\n",
    "\n",
    "def compute_rmse(actual, predicted):\n",
    "  return np.sqrt(np.mean((actual - predicted)**2))\n",
    "\n",
    "def print_rmse(df, rate, name):\n",
    "  print(\"{1} RMSE = {0}\".format(compute_rmse(df['fare_amount'], rate * estimate_distance(df)), name))\n",
    "\n",
    "FEATURES = ['pickuplon','pickuplat','dropofflon','dropofflat','passengers']\n",
    "TARGET = 'fare_amount'\n",
    "columns = list([TARGET])\n",
    "columns.extend(FEATURES) # in CSV, target is the first column, after the features\n",
    "columns.append('key')\n",
    "df_train = pd.read_csv('taxi-train.csv', header = None, names = columns)\n",
    "df_valid = pd.read_csv('taxi-valid.csv', header = None, names = columns)\n",
    "df_test = pd.read_csv('taxi-test.csv', header = None, names = columns)\n",
    "rate = df_train['fare_amount'].mean() / estimate_distance(df_train).mean()\n",
    "print(\"Rate = ${0}/km\".format(rate))\n",
    "print_rmse(df_train, rate, 'Train')\n",
    "print_rmse(df_valid, rate, 'Valid') \n",
    "print_rmse(df_test, rate, 'Test') "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The simple distance-based rule gives us a RMSE of <b>$9.35</b> on the validation dataset.  We have to beat this, of course, but you will find that simple rules of thumb like this can be surprisingly difficult to beat. You don't wnat to set a goal on the test dataset because you want to change the architecture of the network etc. to get the best validation error. Then, you can evaluate ONCE on the test data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Challenge Exercise\n",
    "\n",
    "Let's say that you want to predict whether a Stackoverflow question will be acceptably answered. Using this [public dataset of questions](https://bigquery.cloud.google.com/table/bigquery-public-data:stackoverflow.posts_questions), create a machine learning dataset that you can use for classification.\n",
    "<p>\n",
    "What is a reasonable benchmark for this problem?\n",
    "What features might be useful?\n",
    "<p>\n",
    "If you got the above easily, try this harder problem: you want to predict whether a question will be acceptably answered within 2 days. How would you create the dataset?\n",
    "<p>\n",
    "Hint (highlight to see):\n",
    "<p style='color:white' linkstyle='color:white'> \n",
    "You will need to do a SQL join with the table of [answers]( https://bigquery.cloud.google.com/table/bigquery-public-data:stackoverflow.posts_answers) to determine whether the answer was within 2 days.\n",
    "</p>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Copyright 2018 Google Inc.\n",
    "Licensed under the Apache License, Version 2.0 (the \"License\"); you may not use this file except in compliance with the License. You may obtain a copy of the License at\n",
    "http://www.apache.org/licenses/LICENSE-2.0\n",
    "Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License."
   ]
  }
 ],
 "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": "3.5.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}