geopandas_taxi.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Install Dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!conda install -y -c anaconda -y networkx\n",
    "!conda install -y -c anaconda -y requests\n",
    "!conda install -y -c anaconda -y pandas # Important\n",
    "!conda install -y -c conda-forge -y geopandas\n",
    "!conda install -y -c anaconda -y scipy # Important\n",
    "!conda install -y -c anaconda -y numpy\n",
    "!conda install -y -c anaconda -y scikit-learn # Important Maybe\n",
    "!conda install -y osmnx=0.11.4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "could not import non-essential module osmnx\n"
     ]
    }
   ],
   "source": [
    "# Import the notebook with some useful functions\n",
    "%run util.ipynb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Read Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "test = pd.read_csv('data/processed_test.zip')\n",
    "train = pd.read_csv('data/processed_train.zip')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(790891, 16)\n",
      "(3927323, 16)\n"
     ]
    }
   ],
   "source": [
    "print(test.shape)\n",
    "print(train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Remove everything outside Manhattan area rectangle\n",
    "test = test[(test.pickup_longitude.between(-74.02, -73.91)) &\\\n",
    "            (test.pickup_latitude.between(40.70, 40.86)) & \\\n",
    "            (test.dropoff_longitude.between(-74.02, -73.91)) &\\\n",
    "            (test.dropoff_latitude.between(40.70, 40.86))]\n",
    "train = train[(train.pickup_longitude.between(-74.02, -73.91)) &\\\n",
    "            (train.pickup_latitude.between(40.70, 40.86)) & \\\n",
    "            (train.dropoff_longitude.between(-74.02, -73.91)) &\\\n",
    "            (train.dropoff_latitude.between(40.70, 40.86))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(710982, 16)\n",
      "(3343247, 16)\n"
     ]
    }
   ],
   "source": [
    "print(test.shape)\n",
    "print(train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>Unnamed: 0</th>\n",
       "      <th>dropoff_datetime</th>\n",
       "      <th>dropoff_latitude</th>\n",
       "      <th>dropoff_longitude</th>\n",
       "      <th>id</th>\n",
       "      <th>passenger_count</th>\n",
       "      <th>pickup_datetime</th>\n",
       "      <th>pickup_latitude</th>\n",
       "      <th>pickup_longitude</th>\n",
       "      <th>store_and_fwd_flag</th>\n",
       "      <th>trip_duration</th>\n",
       "      <th>vendor_id</th>\n",
       "      <th>node_lat</th>\n",
       "      <th>node_lng</th>\n",
       "      <th>order</th>\n",
       "      <th>osm_node</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>2016-02-18 23:59:21</td>\n",
       "      <td>40.759827</td>\n",
       "      <td>-73.98983</td>\n",
       "      <td>id3070143</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-02-18 23:50:13</td>\n",
       "      <td>40.750599</td>\n",
       "      <td>-73.968559</td>\n",
       "      <td>N</td>\n",
       "      <td>548</td>\n",
       "      <td>1</td>\n",
       "      <td>40.750751</td>\n",
       "      <td>-73.968402</td>\n",
       "      <td>0</td>\n",
       "      <td>42444055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2016-02-18 23:59:21</td>\n",
       "      <td>40.759827</td>\n",
       "      <td>-73.98983</td>\n",
       "      <td>id3070143</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-02-18 23:50:13</td>\n",
       "      <td>40.750599</td>\n",
       "      <td>-73.968559</td>\n",
       "      <td>N</td>\n",
       "      <td>548</td>\n",
       "      <td>1</td>\n",
       "      <td>40.751748</td>\n",
       "      <td>-73.970804</td>\n",
       "      <td>1</td>\n",
       "      <td>42442947</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>2016-02-18 23:59:21</td>\n",
       "      <td>40.759827</td>\n",
       "      <td>-73.98983</td>\n",
       "      <td>id3070143</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-02-18 23:50:13</td>\n",
       "      <td>40.750599</td>\n",
       "      <td>-73.968559</td>\n",
       "      <td>N</td>\n",
       "      <td>548</td>\n",
       "      <td>1</td>\n",
       "      <td>40.752692</td>\n",
       "      <td>-73.973037</td>\n",
       "      <td>2</td>\n",
       "      <td>42444051</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>2016-02-18 23:59:21</td>\n",
       "      <td>40.759827</td>\n",
       "      <td>-73.98983</td>\n",
       "      <td>id3070143</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-02-18 23:50:13</td>\n",
       "      <td>40.750599</td>\n",
       "      <td>-73.968559</td>\n",
       "      <td>N</td>\n",
       "      <td>548</td>\n",
       "      <td>1</td>\n",
       "      <td>40.753367</td>\n",
       "      <td>-73.974648</td>\n",
       "      <td>3</td>\n",
       "      <td>42444049</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>2016-02-18 23:59:21</td>\n",
       "      <td>40.759827</td>\n",
       "      <td>-73.98983</td>\n",
       "      <td>id3070143</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-02-18 23:50:13</td>\n",
       "      <td>40.750599</td>\n",
       "      <td>-73.968559</td>\n",
       "      <td>N</td>\n",
       "      <td>548</td>\n",
       "      <td>1</td>\n",
       "      <td>40.754349</td>\n",
       "      <td>-73.976974</td>\n",
       "      <td>4</td>\n",
       "      <td>42443810</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0     dropoff_datetime  dropoff_latitude  dropoff_longitude  \\\n",
       "0           0  2016-02-18 23:59:21         40.759827          -73.98983   \n",
       "1           1  2016-02-18 23:59:21         40.759827          -73.98983   \n",
       "2           2  2016-02-18 23:59:21         40.759827          -73.98983   \n",
       "3           3  2016-02-18 23:59:21         40.759827          -73.98983   \n",
       "4           4  2016-02-18 23:59:21         40.759827          -73.98983   \n",
       "\n",
       "          id  passenger_count      pickup_datetime  pickup_latitude  \\\n",
       "0  id3070143                1  2016-02-18 23:50:13        40.750599   \n",
       "1  id3070143                1  2016-02-18 23:50:13        40.750599   \n",
       "2  id3070143                1  2016-02-18 23:50:13        40.750599   \n",
       "3  id3070143                1  2016-02-18 23:50:13        40.750599   \n",
       "4  id3070143                1  2016-02-18 23:50:13        40.750599   \n",
       "\n",
       "   pickup_longitude store_and_fwd_flag  trip_duration  vendor_id   node_lat  \\\n",
       "0        -73.968559                  N            548          1  40.750751   \n",
       "1        -73.968559                  N            548          1  40.751748   \n",
       "2        -73.968559                  N            548          1  40.752692   \n",
       "3        -73.968559                  N            548          1  40.753367   \n",
       "4        -73.968559                  N            548          1  40.754349   \n",
       "\n",
       "    node_lng  order  osm_node  \n",
       "0 -73.968402      0  42444055  \n",
       "1 -73.970804      1  42442947  \n",
       "2 -73.973037      2  42444051  \n",
       "3 -73.974648      3  42444049  \n",
       "4 -73.976974      4  42443810  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Get Graph (map)\n",
    "\n",
    "* Get a graph (nodes and edges) of a preferred location using string address\n",
    "* Plot a random route just as demo\n",
    "* This demo shows how to get the best route (theoretcally) between two coordinates (it is a relatively slow process)\n",
    "* You only need to use this if you want to improve the data we already created as we already saved the best routes in the csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 272.292x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "G = get_graph('Manhattan, New York City, New York, USA')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 272.292x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_route(G, (40.735018, -74.003215), (40.710040, -74.007163))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Example (geoindex)\n",
    "\n",
    "* Here we create a KDTree and demo how to get nearest points from our dataframe with O(1) time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:1: FutureWarning: The pandas.np module is deprecated and will be removed from pandas in a future version. Import numpy directly instead\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    }
   ],
   "source": [
    "# Build tree (we add noise because we want to avoid having points on top of each other)\n",
    "noise = pd.np.random.uniform(0, 0.000001, size=(len(train), 2))\n",
    "tree = KDTree(data=train[['node_lat', 'node_lng']].values + noise)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": true
   },
   "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>Unnamed: 0</th>\n",
       "      <th>dropoff_datetime</th>\n",
       "      <th>dropoff_latitude</th>\n",
       "      <th>dropoff_longitude</th>\n",
       "      <th>id</th>\n",
       "      <th>passenger_count</th>\n",
       "      <th>pickup_datetime</th>\n",
       "      <th>pickup_latitude</th>\n",
       "      <th>pickup_longitude</th>\n",
       "      <th>store_and_fwd_flag</th>\n",
       "      <th>trip_duration</th>\n",
       "      <th>vendor_id</th>\n",
       "      <th>node_lat</th>\n",
       "      <th>node_lng</th>\n",
       "      <th>order</th>\n",
       "      <th>osm_node</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1620548</th>\n",
       "      <td>1620548</td>\n",
       "      <td>2016-01-13 16:30:18</td>\n",
       "      <td>40.734341</td>\n",
       "      <td>-74.008522</td>\n",
       "      <td>id0748042</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-01-13 16:20:39</td>\n",
       "      <td>40.740768</td>\n",
       "      <td>-73.990120</td>\n",
       "      <td>N</td>\n",
       "      <td>579</td>\n",
       "      <td>1</td>\n",
       "      <td>40.734891</td>\n",
       "      <td>-74.003153</td>\n",
       "      <td>17</td>\n",
       "      <td>42452598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3085896</th>\n",
       "      <td>3085896</td>\n",
       "      <td>2016-01-11 14:39:53</td>\n",
       "      <td>40.734604</td>\n",
       "      <td>-74.005981</td>\n",
       "      <td>id1387519</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-01-11 14:32:36</td>\n",
       "      <td>40.735985</td>\n",
       "      <td>-73.993614</td>\n",
       "      <td>N</td>\n",
       "      <td>437</td>\n",
       "      <td>1</td>\n",
       "      <td>40.734891</td>\n",
       "      <td>-74.003153</td>\n",
       "      <td>9</td>\n",
       "      <td>42452598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>380214</th>\n",
       "      <td>380214</td>\n",
       "      <td>2016-01-29 21:41:27</td>\n",
       "      <td>40.734695</td>\n",
       "      <td>-74.004646</td>\n",
       "      <td>id0056842</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-01-29 21:32:39</td>\n",
       "      <td>40.726124</td>\n",
       "      <td>-74.001450</td>\n",
       "      <td>N</td>\n",
       "      <td>528</td>\n",
       "      <td>2</td>\n",
       "      <td>40.734891</td>\n",
       "      <td>-74.003153</td>\n",
       "      <td>17</td>\n",
       "      <td>42452598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>435011</th>\n",
       "      <td>435011</td>\n",
       "      <td>2016-02-12 00:36:45</td>\n",
       "      <td>40.738144</td>\n",
       "      <td>-74.009048</td>\n",
       "      <td>id1777013</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-02-12 00:30:48</td>\n",
       "      <td>40.733734</td>\n",
       "      <td>-74.002335</td>\n",
       "      <td>N</td>\n",
       "      <td>357</td>\n",
       "      <td>2</td>\n",
       "      <td>40.734891</td>\n",
       "      <td>-74.003153</td>\n",
       "      <td>3</td>\n",
       "      <td>42452598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1610027</th>\n",
       "      <td>1610027</td>\n",
       "      <td>2016-02-27 20:47:05</td>\n",
       "      <td>40.740192</td>\n",
       "      <td>-74.005653</td>\n",
       "      <td>id3546252</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-02-27 20:42:35</td>\n",
       "      <td>40.734051</td>\n",
       "      <td>-74.002876</td>\n",
       "      <td>N</td>\n",
       "      <td>270</td>\n",
       "      <td>2</td>\n",
       "      <td>40.734891</td>\n",
       "      <td>-74.003153</td>\n",
       "      <td>1</td>\n",
       "      <td>42452598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2494943</th>\n",
       "      <td>2494943</td>\n",
       "      <td>2016-02-01 21:30:34</td>\n",
       "      <td>40.733326</td>\n",
       "      <td>-74.008400</td>\n",
       "      <td>id1247731</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-02-01 21:24:48</td>\n",
       "      <td>40.736771</td>\n",
       "      <td>-73.997513</td>\n",
       "      <td>N</td>\n",
       "      <td>346</td>\n",
       "      <td>1</td>\n",
       "      <td>40.734891</td>\n",
       "      <td>-74.003153</td>\n",
       "      <td>7</td>\n",
       "      <td>42452598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2756240</th>\n",
       "      <td>2756240</td>\n",
       "      <td>2016-06-23 09:52:01</td>\n",
       "      <td>40.734192</td>\n",
       "      <td>-74.010040</td>\n",
       "      <td>id1145728</td>\n",
       "      <td>2</td>\n",
       "      <td>2016-06-23 09:40:29</td>\n",
       "      <td>40.724884</td>\n",
       "      <td>-73.994331</td>\n",
       "      <td>N</td>\n",
       "      <td>692</td>\n",
       "      <td>1</td>\n",
       "      <td>40.734891</td>\n",
       "      <td>-74.003153</td>\n",
       "      <td>19</td>\n",
       "      <td>42452598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3211168</th>\n",
       "      <td>3211168</td>\n",
       "      <td>2016-04-01 07:13:03</td>\n",
       "      <td>40.733700</td>\n",
       "      <td>-74.006920</td>\n",
       "      <td>id0665117</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-04-01 07:02:43</td>\n",
       "      <td>40.730759</td>\n",
       "      <td>-73.982979</td>\n",
       "      <td>N</td>\n",
       "      <td>620</td>\n",
       "      <td>2</td>\n",
       "      <td>40.734891</td>\n",
       "      <td>-74.003153</td>\n",
       "      <td>14</td>\n",
       "      <td>42452598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1031658</th>\n",
       "      <td>1031658</td>\n",
       "      <td>2016-06-04 19:15:45</td>\n",
       "      <td>40.735500</td>\n",
       "      <td>-74.003357</td>\n",
       "      <td>id2725912</td>\n",
       "      <td>2</td>\n",
       "      <td>2016-06-04 18:55:26</td>\n",
       "      <td>40.726292</td>\n",
       "      <td>-73.980324</td>\n",
       "      <td>N</td>\n",
       "      <td>1219</td>\n",
       "      <td>2</td>\n",
       "      <td>40.734891</td>\n",
       "      <td>-74.003153</td>\n",
       "      <td>20</td>\n",
       "      <td>42452598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3070591</th>\n",
       "      <td>3070591</td>\n",
       "      <td>2016-06-09 21:23:14</td>\n",
       "      <td>40.735424</td>\n",
       "      <td>-74.003349</td>\n",
       "      <td>id1203667</td>\n",
       "      <td>3</td>\n",
       "      <td>2016-06-09 21:14:16</td>\n",
       "      <td>40.729885</td>\n",
       "      <td>-73.991737</td>\n",
       "      <td>N</td>\n",
       "      <td>538</td>\n",
       "      <td>1</td>\n",
       "      <td>40.734891</td>\n",
       "      <td>-74.003153</td>\n",
       "      <td>12</td>\n",
       "      <td>42452598</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>199 rows × 16 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         Unnamed: 0     dropoff_datetime  dropoff_latitude  dropoff_longitude  \\\n",
       "1620548     1620548  2016-01-13 16:30:18         40.734341         -74.008522   \n",
       "3085896     3085896  2016-01-11 14:39:53         40.734604         -74.005981   \n",
       "380214       380214  2016-01-29 21:41:27         40.734695         -74.004646   \n",
       "435011       435011  2016-02-12 00:36:45         40.738144         -74.009048   \n",
       "1610027     1610027  2016-02-27 20:47:05         40.740192         -74.005653   \n",
       "...             ...                  ...               ...                ...   \n",
       "2494943     2494943  2016-02-01 21:30:34         40.733326         -74.008400   \n",
       "2756240     2756240  2016-06-23 09:52:01         40.734192         -74.010040   \n",
       "3211168     3211168  2016-04-01 07:13:03         40.733700         -74.006920   \n",
       "1031658     1031658  2016-06-04 19:15:45         40.735500         -74.003357   \n",
       "3070591     3070591  2016-06-09 21:23:14         40.735424         -74.003349   \n",
       "\n",
       "                id  passenger_count      pickup_datetime  pickup_latitude  \\\n",
       "1620548  id0748042                1  2016-01-13 16:20:39        40.740768   \n",
       "3085896  id1387519                1  2016-01-11 14:32:36        40.735985   \n",
       "380214   id0056842                1  2016-01-29 21:32:39        40.726124   \n",
       "435011   id1777013                1  2016-02-12 00:30:48        40.733734   \n",
       "1610027  id3546252                1  2016-02-27 20:42:35        40.734051   \n",
       "...            ...              ...                  ...              ...   \n",
       "2494943  id1247731                1  2016-02-01 21:24:48        40.736771   \n",
       "2756240  id1145728                2  2016-06-23 09:40:29        40.724884   \n",
       "3211168  id0665117                1  2016-04-01 07:02:43        40.730759   \n",
       "1031658  id2725912                2  2016-06-04 18:55:26        40.726292   \n",
       "3070591  id1203667                3  2016-06-09 21:14:16        40.729885   \n",
       "\n",
       "         pickup_longitude store_and_fwd_flag  trip_duration  vendor_id  \\\n",
       "1620548        -73.990120                  N            579          1   \n",
       "3085896        -73.993614                  N            437          1   \n",
       "380214         -74.001450                  N            528          2   \n",
       "435011         -74.002335                  N            357          2   \n",
       "1610027        -74.002876                  N            270          2   \n",
       "...                   ...                ...            ...        ...   \n",
       "2494943        -73.997513                  N            346          1   \n",
       "2756240        -73.994331                  N            692          1   \n",
       "3211168        -73.982979                  N            620          2   \n",
       "1031658        -73.980324                  N           1219          2   \n",
       "3070591        -73.991737                  N            538          1   \n",
       "\n",
       "          node_lat   node_lng  order  osm_node  \n",
       "1620548  40.734891 -74.003153     17  42452598  \n",
       "3085896  40.734891 -74.003153      9  42452598  \n",
       "380214   40.734891 -74.003153     17  42452598  \n",
       "435011   40.734891 -74.003153      3  42452598  \n",
       "1610027  40.734891 -74.003153      1  42452598  \n",
       "...            ...        ...    ...       ...  \n",
       "2494943  40.734891 -74.003153      7  42452598  \n",
       "2756240  40.734891 -74.003153     19  42452598  \n",
       "3211168  40.734891 -74.003153     14  42452598  \n",
       "1031658  40.734891 -74.003153     20  42452598  \n",
       "3070591  40.734891 -74.003153     12  42452598  \n",
       "\n",
       "[199 rows x 16 columns]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Points within approx 25 meters\n",
    "train.iloc[tree.query_ball_point((40.735018, -74.003215), 0.0003)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Baseline\n",
    "\n",
    "* Here we are predicting that all taxi rides in the test set will take simply the average time in the training set\n",
    "* This is the baseline to beat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "trip_duration    871.235925\n",
       "dtype: float64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calc mean trip duration in the training set\n",
    "mean = train.groupby('id').agg({'trip_duration': 'first'}).mean()\n",
    "test['pred'] = mean['trip_duration']\n",
    "mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0872240148148207"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Evaluate\n",
    "evaluate(test, 'trip_duration', 'pred')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.6"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}