{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.ticker as ticker\n",
    "from IPython.core.display import display, HTML\n",
    "import os\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "import missingno as msno"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>.container { width:100% !important; }</style>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Notebook Styling \n",
    "sns.set()\n",
    "pd.options.display.max_columns = None\n",
    "display(HTML(\"<style>.container { width:100% !important; }</style>\"))\n",
    "pd.set_option('display.float_format',lambda x: '%.5f' % x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Problem Description\n",
    "\n",
    "Your goal is to predict the operating condition of a waterpoint for each record in the dataset. You are provided the following set of information about the waterpoints:\n",
    "\n",
    "| Variable | Variable Description |\n",
    "| --- | :--- |\n",
    "| amount_tsh | Total static head (amount water available to waterpoint) |\n",
    "| date_recorded | The date the row was entered |\n",
    "| funder | Who funded the well |\n",
    "| gps_height | Altitude of the well |\n",
    "| installer | Organization that installed the well |\n",
    "| longitude | GPS coordinate |\n",
    "| latitude | GPS coordinate |\n",
    "| wpt_name | Name of the waterpoint if there is one |\n",
    "| num_private | <blank> |\n",
    "| basin | Geographic water basin |\n",
    "| subvillage | Geographic location |\n",
    "| region | Geographic location |\n",
    "| region_code | Geographic location (coded) |\n",
    "| district_code | Geographic location (coded) |\n",
    "| lga | Geographic location |\n",
    "| ward | Geographic location |\n",
    "| population | Population around the well |\n",
    "| public_meeting | True/False |\n",
    "| recorded_by | Group entering this row of data |\n",
    "| scheme_management | Who operates the waterpoint |\n",
    "| scheme_name | Who operates the waterpoint |\n",
    "| permit | If the waterpoint is permitted |\n",
    "| construction_year | Year the waterpoint was constructed |\n",
    "| extraction_type | The kind of extraction the waterpoint uses |\n",
    "| extraction_type_group | The kind of extraction the waterpoint uses |\n",
    "| extraction_type_class | The kind of extraction the waterpoint uses |\n",
    "| management | How the waterpoint is managed |\n",
    "| management_group | How the waterpoint is managed |\n",
    "| payment | What the water costs |\n",
    "| payment_type | What the water costs |\n",
    "| water_quality | The quality of the water |\n",
    "| quality_group | The quality of the water |\n",
    "| quantity | The quantity of water |\n",
    "| quantity_group | The quantity of water |\n",
    "| source | The source of the water |\n",
    "| source_type | The source of the water |\n",
    "| source_class | The source of the water |\n",
    "| waterpoint_type | The kind of waterpoint |\n",
    "| waterpoint_type_group | The kind of waterpoint |"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>amount_tsh</th>\n",
       "      <th>date_recorded</th>\n",
       "      <th>funder</th>\n",
       "      <th>gps_height</th>\n",
       "      <th>installer</th>\n",
       "      <th>longitude</th>\n",
       "      <th>latitude</th>\n",
       "      <th>wpt_name</th>\n",
       "      <th>num_private</th>\n",
       "      <th>basin</th>\n",
       "      <th>subvillage</th>\n",
       "      <th>region</th>\n",
       "      <th>region_code</th>\n",
       "      <th>district_code</th>\n",
       "      <th>lga</th>\n",
       "      <th>ward</th>\n",
       "      <th>population</th>\n",
       "      <th>public_meeting</th>\n",
       "      <th>recorded_by</th>\n",
       "      <th>scheme_management</th>\n",
       "      <th>scheme_name</th>\n",
       "      <th>permit</th>\n",
       "      <th>construction_year</th>\n",
       "      <th>extraction_type</th>\n",
       "      <th>extraction_type_group</th>\n",
       "      <th>extraction_type_class</th>\n",
       "      <th>management</th>\n",
       "      <th>management_group</th>\n",
       "      <th>payment</th>\n",
       "      <th>payment_type</th>\n",
       "      <th>water_quality</th>\n",
       "      <th>quality_group</th>\n",
       "      <th>quantity</th>\n",
       "      <th>quantity_group</th>\n",
       "      <th>source</th>\n",
       "      <th>source_type</th>\n",
       "      <th>source_class</th>\n",
       "      <th>waterpoint_type</th>\n",
       "      <th>waterpoint_type_group</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>69572</th>\n",
       "      <td>6000.00000</td>\n",
       "      <td>2011-03-14</td>\n",
       "      <td>Roman</td>\n",
       "      <td>1390</td>\n",
       "      <td>Roman</td>\n",
       "      <td>34.93809</td>\n",
       "      <td>-9.85632</td>\n",
       "      <td>none</td>\n",
       "      <td>0</td>\n",
       "      <td>Lake Nyasa</td>\n",
       "      <td>Mnyusi B</td>\n",
       "      <td>Iringa</td>\n",
       "      <td>11</td>\n",
       "      <td>5</td>\n",
       "      <td>Ludewa</td>\n",
       "      <td>Mundindi</td>\n",
       "      <td>109</td>\n",
       "      <td>True</td>\n",
       "      <td>GeoData Consultants Ltd</td>\n",
       "      <td>VWC</td>\n",
       "      <td>Roman</td>\n",
       "      <td>False</td>\n",
       "      <td>1999</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>vwc</td>\n",
       "      <td>user-group</td>\n",
       "      <td>pay annually</td>\n",
       "      <td>annually</td>\n",
       "      <td>soft</td>\n",
       "      <td>good</td>\n",
       "      <td>enough</td>\n",
       "      <td>enough</td>\n",
       "      <td>spring</td>\n",
       "      <td>spring</td>\n",
       "      <td>groundwater</td>\n",
       "      <td>communal standpipe</td>\n",
       "      <td>communal standpipe</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8776</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>2013-03-06</td>\n",
       "      <td>Grumeti</td>\n",
       "      <td>1399</td>\n",
       "      <td>GRUMETI</td>\n",
       "      <td>34.69877</td>\n",
       "      <td>-2.14747</td>\n",
       "      <td>Zahanati</td>\n",
       "      <td>0</td>\n",
       "      <td>Lake Victoria</td>\n",
       "      <td>Nyamara</td>\n",
       "      <td>Mara</td>\n",
       "      <td>20</td>\n",
       "      <td>2</td>\n",
       "      <td>Serengeti</td>\n",
       "      <td>Natta</td>\n",
       "      <td>280</td>\n",
       "      <td>NaN</td>\n",
       "      <td>GeoData Consultants Ltd</td>\n",
       "      <td>Other</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>2010</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>wug</td>\n",
       "      <td>user-group</td>\n",
       "      <td>never pay</td>\n",
       "      <td>never pay</td>\n",
       "      <td>soft</td>\n",
       "      <td>good</td>\n",
       "      <td>insufficient</td>\n",
       "      <td>insufficient</td>\n",
       "      <td>rainwater harvesting</td>\n",
       "      <td>rainwater harvesting</td>\n",
       "      <td>surface</td>\n",
       "      <td>communal standpipe</td>\n",
       "      <td>communal standpipe</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34310</th>\n",
       "      <td>25.00000</td>\n",
       "      <td>2013-02-25</td>\n",
       "      <td>Lottery Club</td>\n",
       "      <td>686</td>\n",
       "      <td>World vision</td>\n",
       "      <td>37.46066</td>\n",
       "      <td>-3.82133</td>\n",
       "      <td>Kwa Mahundi</td>\n",
       "      <td>0</td>\n",
       "      <td>Pangani</td>\n",
       "      <td>Majengo</td>\n",
       "      <td>Manyara</td>\n",
       "      <td>21</td>\n",
       "      <td>4</td>\n",
       "      <td>Simanjiro</td>\n",
       "      <td>Ngorika</td>\n",
       "      <td>250</td>\n",
       "      <td>True</td>\n",
       "      <td>GeoData Consultants Ltd</td>\n",
       "      <td>VWC</td>\n",
       "      <td>Nyumba ya mungu pipe scheme</td>\n",
       "      <td>True</td>\n",
       "      <td>2009</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>vwc</td>\n",
       "      <td>user-group</td>\n",
       "      <td>pay per bucket</td>\n",
       "      <td>per bucket</td>\n",
       "      <td>soft</td>\n",
       "      <td>good</td>\n",
       "      <td>enough</td>\n",
       "      <td>enough</td>\n",
       "      <td>dam</td>\n",
       "      <td>dam</td>\n",
       "      <td>surface</td>\n",
       "      <td>communal standpipe multiple</td>\n",
       "      <td>communal standpipe</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67743</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>2013-01-28</td>\n",
       "      <td>Unicef</td>\n",
       "      <td>263</td>\n",
       "      <td>UNICEF</td>\n",
       "      <td>38.48616</td>\n",
       "      <td>-11.15530</td>\n",
       "      <td>Zahanati Ya Nanyumbu</td>\n",
       "      <td>0</td>\n",
       "      <td>Ruvuma / Southern Coast</td>\n",
       "      <td>Mahakamani</td>\n",
       "      <td>Mtwara</td>\n",
       "      <td>90</td>\n",
       "      <td>63</td>\n",
       "      <td>Nanyumbu</td>\n",
       "      <td>Nanyumbu</td>\n",
       "      <td>58</td>\n",
       "      <td>True</td>\n",
       "      <td>GeoData Consultants Ltd</td>\n",
       "      <td>VWC</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>1986</td>\n",
       "      <td>submersible</td>\n",
       "      <td>submersible</td>\n",
       "      <td>submersible</td>\n",
       "      <td>vwc</td>\n",
       "      <td>user-group</td>\n",
       "      <td>never pay</td>\n",
       "      <td>never pay</td>\n",
       "      <td>soft</td>\n",
       "      <td>good</td>\n",
       "      <td>dry</td>\n",
       "      <td>dry</td>\n",
       "      <td>machine dbh</td>\n",
       "      <td>borehole</td>\n",
       "      <td>groundwater</td>\n",
       "      <td>communal standpipe multiple</td>\n",
       "      <td>communal standpipe</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19728</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>2011-07-13</td>\n",
       "      <td>Action In A</td>\n",
       "      <td>0</td>\n",
       "      <td>Artisan</td>\n",
       "      <td>31.13085</td>\n",
       "      <td>-1.82536</td>\n",
       "      <td>Shuleni</td>\n",
       "      <td>0</td>\n",
       "      <td>Lake Victoria</td>\n",
       "      <td>Kyanyamisa</td>\n",
       "      <td>Kagera</td>\n",
       "      <td>18</td>\n",
       "      <td>1</td>\n",
       "      <td>Karagwe</td>\n",
       "      <td>Nyakasimbi</td>\n",
       "      <td>0</td>\n",
       "      <td>True</td>\n",
       "      <td>GeoData Consultants Ltd</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>other</td>\n",
       "      <td>other</td>\n",
       "      <td>never pay</td>\n",
       "      <td>never pay</td>\n",
       "      <td>soft</td>\n",
       "      <td>good</td>\n",
       "      <td>seasonal</td>\n",
       "      <td>seasonal</td>\n",
       "      <td>rainwater harvesting</td>\n",
       "      <td>rainwater harvesting</td>\n",
       "      <td>surface</td>\n",
       "      <td>communal standpipe</td>\n",
       "      <td>communal standpipe</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       amount_tsh date_recorded        funder  gps_height     installer  \\\n",
       "id                                                                        \n",
       "69572  6000.00000    2011-03-14         Roman        1390         Roman   \n",
       "8776      0.00000    2013-03-06       Grumeti        1399       GRUMETI   \n",
       "34310    25.00000    2013-02-25  Lottery Club         686  World vision   \n",
       "67743     0.00000    2013-01-28        Unicef         263        UNICEF   \n",
       "19728     0.00000    2011-07-13   Action In A           0       Artisan   \n",
       "\n",
       "       longitude  latitude              wpt_name  num_private  \\\n",
       "id                                                              \n",
       "69572   34.93809  -9.85632                  none            0   \n",
       "8776    34.69877  -2.14747              Zahanati            0   \n",
       "34310   37.46066  -3.82133           Kwa Mahundi            0   \n",
       "67743   38.48616 -11.15530  Zahanati Ya Nanyumbu            0   \n",
       "19728   31.13085  -1.82536               Shuleni            0   \n",
       "\n",
       "                         basin  subvillage   region  region_code  \\\n",
       "id                                                                 \n",
       "69572               Lake Nyasa    Mnyusi B   Iringa           11   \n",
       "8776             Lake Victoria     Nyamara     Mara           20   \n",
       "34310                  Pangani     Majengo  Manyara           21   \n",
       "67743  Ruvuma / Southern Coast  Mahakamani   Mtwara           90   \n",
       "19728            Lake Victoria  Kyanyamisa   Kagera           18   \n",
       "\n",
       "       district_code        lga        ward  population public_meeting  \\\n",
       "id                                                                       \n",
       "69572              5     Ludewa    Mundindi         109           True   \n",
       "8776               2  Serengeti       Natta         280            NaN   \n",
       "34310              4  Simanjiro     Ngorika         250           True   \n",
       "67743             63   Nanyumbu    Nanyumbu          58           True   \n",
       "19728              1    Karagwe  Nyakasimbi           0           True   \n",
       "\n",
       "                   recorded_by scheme_management                  scheme_name  \\\n",
       "id                                                                              \n",
       "69572  GeoData Consultants Ltd               VWC                        Roman   \n",
       "8776   GeoData Consultants Ltd             Other                          NaN   \n",
       "34310  GeoData Consultants Ltd               VWC  Nyumba ya mungu pipe scheme   \n",
       "67743  GeoData Consultants Ltd               VWC                          NaN   \n",
       "19728  GeoData Consultants Ltd               NaN                          NaN   \n",
       "\n",
       "      permit  construction_year extraction_type extraction_type_group  \\\n",
       "id                                                                      \n",
       "69572  False               1999         gravity               gravity   \n",
       "8776    True               2010         gravity               gravity   \n",
       "34310   True               2009         gravity               gravity   \n",
       "67743   True               1986     submersible           submersible   \n",
       "19728   True                  0         gravity               gravity   \n",
       "\n",
       "      extraction_type_class management management_group         payment  \\\n",
       "id                                                                        \n",
       "69572               gravity        vwc       user-group    pay annually   \n",
       "8776                gravity        wug       user-group       never pay   \n",
       "34310               gravity        vwc       user-group  pay per bucket   \n",
       "67743           submersible        vwc       user-group       never pay   \n",
       "19728               gravity      other            other       never pay   \n",
       "\n",
       "      payment_type water_quality quality_group      quantity quantity_group  \\\n",
       "id                                                                            \n",
       "69572     annually          soft          good        enough         enough   \n",
       "8776     never pay          soft          good  insufficient   insufficient   \n",
       "34310   per bucket          soft          good        enough         enough   \n",
       "67743    never pay          soft          good           dry            dry   \n",
       "19728    never pay          soft          good      seasonal       seasonal   \n",
       "\n",
       "                     source           source_type source_class  \\\n",
       "id                                                               \n",
       "69572                spring                spring  groundwater   \n",
       "8776   rainwater harvesting  rainwater harvesting      surface   \n",
       "34310                   dam                   dam      surface   \n",
       "67743           machine dbh              borehole  groundwater   \n",
       "19728  rainwater harvesting  rainwater harvesting      surface   \n",
       "\n",
       "                   waterpoint_type waterpoint_type_group  \n",
       "id                                                        \n",
       "69572           communal standpipe    communal standpipe  \n",
       "8776            communal standpipe    communal standpipe  \n",
       "34310  communal standpipe multiple    communal standpipe  \n",
       "67743  communal standpipe multiple    communal standpipe  \n",
       "19728           communal standpipe    communal standpipe  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "CSV_PATH = os.path.join('data', 'water_pumps', 'training_vals.csv')\n",
    "X_train = pd.read_csv(CSV_PATH, encoding='latin1', index_col='id', parse_dates=['date_recorded']) \n",
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# are all of the indices unique?\n",
    "bool(len(X_train.index.unique()) == X_train.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(59400, 40)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>status_group</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>69572</th>\n",
       "      <td>functional</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8776</th>\n",
       "      <td>functional</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34310</th>\n",
       "      <td>functional</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67743</th>\n",
       "      <td>non functional</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19728</th>\n",
       "      <td>functional</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         status_group\n",
       "id                   \n",
       "69572      functional\n",
       "8776       functional\n",
       "34310      functional\n",
       "67743  non functional\n",
       "19728      functional"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "CSV_PATH = os.path.join('data', 'water_pumps', 'training_labels.csv')\n",
    "y_train = pd.read_csv(CSV_PATH, encoding='latin1', index_col='id') \n",
    "y_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "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>amount_tsh</th>\n",
       "      <th>date_recorded</th>\n",
       "      <th>funder</th>\n",
       "      <th>gps_height</th>\n",
       "      <th>installer</th>\n",
       "      <th>longitude</th>\n",
       "      <th>latitude</th>\n",
       "      <th>wpt_name</th>\n",
       "      <th>num_private</th>\n",
       "      <th>basin</th>\n",
       "      <th>subvillage</th>\n",
       "      <th>region</th>\n",
       "      <th>region_code</th>\n",
       "      <th>district_code</th>\n",
       "      <th>lga</th>\n",
       "      <th>ward</th>\n",
       "      <th>population</th>\n",
       "      <th>public_meeting</th>\n",
       "      <th>recorded_by</th>\n",
       "      <th>scheme_management</th>\n",
       "      <th>scheme_name</th>\n",
       "      <th>permit</th>\n",
       "      <th>construction_year</th>\n",
       "      <th>extraction_type</th>\n",
       "      <th>extraction_type_group</th>\n",
       "      <th>extraction_type_class</th>\n",
       "      <th>management</th>\n",
       "      <th>management_group</th>\n",
       "      <th>payment</th>\n",
       "      <th>payment_type</th>\n",
       "      <th>water_quality</th>\n",
       "      <th>quality_group</th>\n",
       "      <th>quantity</th>\n",
       "      <th>quantity_group</th>\n",
       "      <th>source</th>\n",
       "      <th>source_type</th>\n",
       "      <th>source_class</th>\n",
       "      <th>waterpoint_type</th>\n",
       "      <th>waterpoint_type_group</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>50785</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>2013-02-04</td>\n",
       "      <td>Dmdd</td>\n",
       "      <td>1996</td>\n",
       "      <td>DMDD</td>\n",
       "      <td>35.29080</td>\n",
       "      <td>-4.05970</td>\n",
       "      <td>Dinamu Secondary School</td>\n",
       "      <td>0</td>\n",
       "      <td>Internal</td>\n",
       "      <td>Magoma</td>\n",
       "      <td>Manyara</td>\n",
       "      <td>21</td>\n",
       "      <td>3</td>\n",
       "      <td>Mbulu</td>\n",
       "      <td>Bashay</td>\n",
       "      <td>321</td>\n",
       "      <td>True</td>\n",
       "      <td>GeoData Consultants Ltd</td>\n",
       "      <td>Parastatal</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>2012</td>\n",
       "      <td>other</td>\n",
       "      <td>other</td>\n",
       "      <td>other</td>\n",
       "      <td>parastatal</td>\n",
       "      <td>parastatal</td>\n",
       "      <td>never pay</td>\n",
       "      <td>never pay</td>\n",
       "      <td>soft</td>\n",
       "      <td>good</td>\n",
       "      <td>seasonal</td>\n",
       "      <td>seasonal</td>\n",
       "      <td>rainwater harvesting</td>\n",
       "      <td>rainwater harvesting</td>\n",
       "      <td>surface</td>\n",
       "      <td>other</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51630</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>2013-02-04</td>\n",
       "      <td>Government Of Tanzania</td>\n",
       "      <td>1569</td>\n",
       "      <td>DWE</td>\n",
       "      <td>36.65671</td>\n",
       "      <td>-3.30921</td>\n",
       "      <td>Kimnyak</td>\n",
       "      <td>0</td>\n",
       "      <td>Pangani</td>\n",
       "      <td>Kimnyak</td>\n",
       "      <td>Arusha</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>Arusha Rural</td>\n",
       "      <td>Kimnyaki</td>\n",
       "      <td>300</td>\n",
       "      <td>True</td>\n",
       "      <td>GeoData Consultants Ltd</td>\n",
       "      <td>VWC</td>\n",
       "      <td>TPRI pipe line</td>\n",
       "      <td>True</td>\n",
       "      <td>2000</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>vwc</td>\n",
       "      <td>user-group</td>\n",
       "      <td>never pay</td>\n",
       "      <td>never pay</td>\n",
       "      <td>soft</td>\n",
       "      <td>good</td>\n",
       "      <td>insufficient</td>\n",
       "      <td>insufficient</td>\n",
       "      <td>spring</td>\n",
       "      <td>spring</td>\n",
       "      <td>groundwater</td>\n",
       "      <td>communal standpipe</td>\n",
       "      <td>communal standpipe</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17168</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>2013-02-01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1567</td>\n",
       "      <td>NaN</td>\n",
       "      <td>34.76786</td>\n",
       "      <td>-5.00434</td>\n",
       "      <td>Puma Secondary</td>\n",
       "      <td>0</td>\n",
       "      <td>Internal</td>\n",
       "      <td>Msatu</td>\n",
       "      <td>Singida</td>\n",
       "      <td>13</td>\n",
       "      <td>2</td>\n",
       "      <td>Singida Rural</td>\n",
       "      <td>Puma</td>\n",
       "      <td>500</td>\n",
       "      <td>True</td>\n",
       "      <td>GeoData Consultants Ltd</td>\n",
       "      <td>VWC</td>\n",
       "      <td>P</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2010</td>\n",
       "      <td>other</td>\n",
       "      <td>other</td>\n",
       "      <td>other</td>\n",
       "      <td>vwc</td>\n",
       "      <td>user-group</td>\n",
       "      <td>never pay</td>\n",
       "      <td>never pay</td>\n",
       "      <td>soft</td>\n",
       "      <td>good</td>\n",
       "      <td>insufficient</td>\n",
       "      <td>insufficient</td>\n",
       "      <td>rainwater harvesting</td>\n",
       "      <td>rainwater harvesting</td>\n",
       "      <td>surface</td>\n",
       "      <td>other</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45559</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>2013-01-22</td>\n",
       "      <td>Finn Water</td>\n",
       "      <td>267</td>\n",
       "      <td>FINN WATER</td>\n",
       "      <td>38.05805</td>\n",
       "      <td>-9.41867</td>\n",
       "      <td>Kwa Mzee Pange</td>\n",
       "      <td>0</td>\n",
       "      <td>Ruvuma / Southern Coast</td>\n",
       "      <td>Kipindimbi</td>\n",
       "      <td>Lindi</td>\n",
       "      <td>80</td>\n",
       "      <td>43</td>\n",
       "      <td>Liwale</td>\n",
       "      <td>Mkutano</td>\n",
       "      <td>250</td>\n",
       "      <td>NaN</td>\n",
       "      <td>GeoData Consultants Ltd</td>\n",
       "      <td>VWC</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>1987</td>\n",
       "      <td>other</td>\n",
       "      <td>other</td>\n",
       "      <td>other</td>\n",
       "      <td>vwc</td>\n",
       "      <td>user-group</td>\n",
       "      <td>unknown</td>\n",
       "      <td>unknown</td>\n",
       "      <td>soft</td>\n",
       "      <td>good</td>\n",
       "      <td>dry</td>\n",
       "      <td>dry</td>\n",
       "      <td>shallow well</td>\n",
       "      <td>shallow well</td>\n",
       "      <td>groundwater</td>\n",
       "      <td>other</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49871</th>\n",
       "      <td>500.00000</td>\n",
       "      <td>2013-03-27</td>\n",
       "      <td>Bruder</td>\n",
       "      <td>1260</td>\n",
       "      <td>BRUDER</td>\n",
       "      <td>35.00612</td>\n",
       "      <td>-10.95041</td>\n",
       "      <td>Kwa Mzee Turuka</td>\n",
       "      <td>0</td>\n",
       "      <td>Ruvuma / Southern Coast</td>\n",
       "      <td>Losonga</td>\n",
       "      <td>Ruvuma</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "      <td>Mbinga</td>\n",
       "      <td>Mbinga Urban</td>\n",
       "      <td>60</td>\n",
       "      <td>NaN</td>\n",
       "      <td>GeoData Consultants Ltd</td>\n",
       "      <td>Water Board</td>\n",
       "      <td>BRUDER</td>\n",
       "      <td>True</td>\n",
       "      <td>2000</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>gravity</td>\n",
       "      <td>water board</td>\n",
       "      <td>user-group</td>\n",
       "      <td>pay monthly</td>\n",
       "      <td>monthly</td>\n",
       "      <td>soft</td>\n",
       "      <td>good</td>\n",
       "      <td>enough</td>\n",
       "      <td>enough</td>\n",
       "      <td>spring</td>\n",
       "      <td>spring</td>\n",
       "      <td>groundwater</td>\n",
       "      <td>communal standpipe</td>\n",
       "      <td>communal standpipe</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       amount_tsh date_recorded                  funder  gps_height  \\\n",
       "id                                                                    \n",
       "50785     0.00000    2013-02-04                    Dmdd        1996   \n",
       "51630     0.00000    2013-02-04  Government Of Tanzania        1569   \n",
       "17168     0.00000    2013-02-01                     NaN        1567   \n",
       "45559     0.00000    2013-01-22              Finn Water         267   \n",
       "49871   500.00000    2013-03-27                  Bruder        1260   \n",
       "\n",
       "        installer  longitude  latitude                 wpt_name  num_private  \\\n",
       "id                                                                             \n",
       "50785        DMDD   35.29080  -4.05970  Dinamu Secondary School            0   \n",
       "51630         DWE   36.65671  -3.30921                  Kimnyak            0   \n",
       "17168         NaN   34.76786  -5.00434           Puma Secondary            0   \n",
       "45559  FINN WATER   38.05805  -9.41867           Kwa Mzee Pange            0   \n",
       "49871      BRUDER   35.00612 -10.95041          Kwa Mzee Turuka            0   \n",
       "\n",
       "                         basin  subvillage   region  region_code  \\\n",
       "id                                                                 \n",
       "50785                 Internal      Magoma  Manyara           21   \n",
       "51630                  Pangani     Kimnyak   Arusha            2   \n",
       "17168                 Internal       Msatu  Singida           13   \n",
       "45559  Ruvuma / Southern Coast  Kipindimbi    Lindi           80   \n",
       "49871  Ruvuma / Southern Coast     Losonga   Ruvuma           10   \n",
       "\n",
       "       district_code            lga          ward  population public_meeting  \\\n",
       "id                                                                             \n",
       "50785              3          Mbulu        Bashay         321           True   \n",
       "51630              2   Arusha Rural      Kimnyaki         300           True   \n",
       "17168              2  Singida Rural          Puma         500           True   \n",
       "45559             43         Liwale       Mkutano         250            NaN   \n",
       "49871              3         Mbinga  Mbinga Urban          60            NaN   \n",
       "\n",
       "                   recorded_by scheme_management     scheme_name permit  \\\n",
       "id                                                                        \n",
       "50785  GeoData Consultants Ltd        Parastatal             NaN   True   \n",
       "51630  GeoData Consultants Ltd               VWC  TPRI pipe line   True   \n",
       "17168  GeoData Consultants Ltd               VWC               P    NaN   \n",
       "45559  GeoData Consultants Ltd               VWC             NaN   True   \n",
       "49871  GeoData Consultants Ltd       Water Board          BRUDER   True   \n",
       "\n",
       "       construction_year extraction_type extraction_type_group  \\\n",
       "id                                                               \n",
       "50785               2012           other                 other   \n",
       "51630               2000         gravity               gravity   \n",
       "17168               2010           other                 other   \n",
       "45559               1987           other                 other   \n",
       "49871               2000         gravity               gravity   \n",
       "\n",
       "      extraction_type_class   management management_group      payment  \\\n",
       "id                                                                       \n",
       "50785                 other   parastatal       parastatal    never pay   \n",
       "51630               gravity          vwc       user-group    never pay   \n",
       "17168                 other          vwc       user-group    never pay   \n",
       "45559                 other          vwc       user-group      unknown   \n",
       "49871               gravity  water board       user-group  pay monthly   \n",
       "\n",
       "      payment_type water_quality quality_group      quantity quantity_group  \\\n",
       "id                                                                            \n",
       "50785    never pay          soft          good      seasonal       seasonal   \n",
       "51630    never pay          soft          good  insufficient   insufficient   \n",
       "17168    never pay          soft          good  insufficient   insufficient   \n",
       "45559      unknown          soft          good           dry            dry   \n",
       "49871      monthly          soft          good        enough         enough   \n",
       "\n",
       "                     source           source_type source_class  \\\n",
       "id                                                               \n",
       "50785  rainwater harvesting  rainwater harvesting      surface   \n",
       "51630                spring                spring  groundwater   \n",
       "17168  rainwater harvesting  rainwater harvesting      surface   \n",
       "45559          shallow well          shallow well  groundwater   \n",
       "49871                spring                spring  groundwater   \n",
       "\n",
       "          waterpoint_type waterpoint_type_group  \n",
       "id                                               \n",
       "50785               other                 other  \n",
       "51630  communal standpipe    communal standpipe  \n",
       "17168               other                 other  \n",
       "45559               other                 other  \n",
       "49871  communal standpipe    communal standpipe  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "CSV_PATH = os.path.join('data', 'water_pumps', 'test_vals.csv')\n",
    "X_test = pd.read_csv(CSV_PATH, encoding='latin1', index_col='id', parse_dates=['date_recorded']) \n",
    "X_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 59400 entries, 69572 to 26348\n",
      "Data columns (total 39 columns):\n",
      "amount_tsh               59400 non-null float64\n",
      "date_recorded            59400 non-null datetime64[ns]\n",
      "funder                   55765 non-null object\n",
      "gps_height               59400 non-null int64\n",
      "installer                55745 non-null object\n",
      "longitude                59400 non-null float64\n",
      "latitude                 59400 non-null float64\n",
      "wpt_name                 59400 non-null object\n",
      "num_private              59400 non-null int64\n",
      "basin                    59400 non-null object\n",
      "subvillage               59029 non-null object\n",
      "region                   59400 non-null object\n",
      "region_code              59400 non-null int64\n",
      "district_code            59400 non-null int64\n",
      "lga                      59400 non-null object\n",
      "ward                     59400 non-null object\n",
      "population               59400 non-null int64\n",
      "public_meeting           56066 non-null object\n",
      "recorded_by              59400 non-null object\n",
      "scheme_management        55523 non-null object\n",
      "scheme_name              31234 non-null object\n",
      "permit                   56344 non-null object\n",
      "construction_year        59400 non-null int64\n",
      "extraction_type          59400 non-null object\n",
      "extraction_type_group    59400 non-null object\n",
      "extraction_type_class    59400 non-null object\n",
      "management               59400 non-null object\n",
      "management_group         59400 non-null object\n",
      "payment                  59400 non-null object\n",
      "payment_type             59400 non-null object\n",
      "water_quality            59400 non-null object\n",
      "quality_group            59400 non-null object\n",
      "quantity                 59400 non-null object\n",
      "quantity_group           59400 non-null object\n",
      "source                   59400 non-null object\n",
      "source_type              59400 non-null object\n",
      "source_class             59400 non-null object\n",
      "waterpoint_type          59400 non-null object\n",
      "waterpoint_type_group    59400 non-null object\n",
      "dtypes: datetime64[ns](1), float64(3), int64(6), object(29)\n",
      "memory usage: 18.1+ MB\n"
     ]
    }
   ],
   "source": [
    "X_train.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Missing Data\n",
    "\n",
    "From the info() printout above, we see that a few of the features are missing some data, and from the plot of missing data below, we see that"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1bf029ea278>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1bf7235b278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "msno.matrix(X_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Possible Categorical Features (features with the Object datatype)\n",
    "Features with the Object datatype must be handled before we attempt to fit a model using the sklearn library. These features could be categorical features, or they could numerical features that require some preprocessing, or there could be some other issue. If features are actually categorical and contain new or important information, we'll have to convert them into dummy variables, and that could quickly cause the number of features to explode, which would promote overfitting. Let's take a look at the number of unique values for each feature."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feat:                funder, n_unique vals train:  1897,  n_unique vals test:   980\n",
      "Feat:             installer, n_unique vals train:  2145,  n_unique vals test:  1091\n",
      "Feat:              wpt_name, n_unique vals train: 37400,  n_unique vals test: 10840\n",
      "Feat:                 basin, n_unique vals train:     9,  n_unique vals test:     9\n",
      "Feat:            subvillage, n_unique vals train: 19287,  n_unique vals test:  8443\n",
      "Feat:                region, n_unique vals train:    21,  n_unique vals test:    21\n",
      "Feat:                   lga, n_unique vals train:   125,  n_unique vals test:   125\n",
      "Feat:                  ward, n_unique vals train:  2092,  n_unique vals test:  1959\n",
      "Feat:        public_meeting, n_unique vals train:     2,  n_unique vals test:     2\n",
      "Feat:           recorded_by, n_unique vals train:     1,  n_unique vals test:     1\n",
      "Feat:     scheme_management, n_unique vals train:    12,  n_unique vals test:    11\n",
      "X_test[feat].unique(): ['Parastatal' 'VWC' 'Water Board' nan 'Other' 'SWC' 'WUG' 'WUA'\n",
      " 'Water authority' 'Company' 'Private operator' 'Trust']\n",
      "X_train[feat].unique(): ['VWC' 'Other' nan 'Private operator' 'WUG' 'Water Board' 'WUA'\n",
      " 'Water authority' 'Company' 'Parastatal' 'Trust' 'SWC' 'None']\n",
      "Feat:           scheme_name, n_unique vals train:  2696,  n_unique vals test:  1789\n",
      "Feat:                permit, n_unique vals train:     2,  n_unique vals test:     2\n",
      "Feat:       extraction_type, n_unique vals train:    18,  n_unique vals test:    17\n",
      "X_test[feat].unique(): ['other' 'gravity' 'india mark ii' 'submersible' 'mono' 'nira/tanira'\n",
      " 'afridev' 'swn 80' 'ksb' 'climax' 'other - rope pump' 'cemo'\n",
      " 'india mark iii' 'other - swn 81' 'other - play pump' 'windmill' 'walimi']\n",
      "X_train[feat].unique(): ['gravity' 'submersible' 'swn 80' 'nira/tanira' 'india mark ii' 'other'\n",
      " 'ksb' 'mono' 'windmill' 'afridev' 'other - rope pump' 'india mark iii'\n",
      " 'other - swn 81' 'other - play pump' 'cemo' 'climax' 'walimi'\n",
      " 'other - mkulima/shinyanga']\n",
      "Feat: extraction_type_group, n_unique vals train:    13,  n_unique vals test:    13\n",
      "Feat: extraction_type_class, n_unique vals train:     7,  n_unique vals test:     7\n",
      "Feat:            management, n_unique vals train:    12,  n_unique vals test:    12\n",
      "Feat:      management_group, n_unique vals train:     5,  n_unique vals test:     5\n",
      "Feat:               payment, n_unique vals train:     7,  n_unique vals test:     7\n",
      "Feat:          payment_type, n_unique vals train:     7,  n_unique vals test:     7\n",
      "Feat:         water_quality, n_unique vals train:     8,  n_unique vals test:     8\n",
      "Feat:         quality_group, n_unique vals train:     6,  n_unique vals test:     6\n",
      "Feat:              quantity, n_unique vals train:     5,  n_unique vals test:     5\n",
      "Feat:        quantity_group, n_unique vals train:     5,  n_unique vals test:     5\n",
      "Feat:                source, n_unique vals train:    10,  n_unique vals test:    10\n",
      "Feat:           source_type, n_unique vals train:     7,  n_unique vals test:     7\n",
      "Feat:          source_class, n_unique vals train:     3,  n_unique vals test:     3\n",
      "Feat:       waterpoint_type, n_unique vals train:     7,  n_unique vals test:     7\n",
      "Feat: waterpoint_type_group, n_unique vals train:     6,  n_unique vals test:     6\n"
     ]
    }
   ],
   "source": [
    "for feat in X_train.columns:\n",
    "    if X_train[feat].dtypes == object:\n",
    "        print('Feat: {:>21s}, n_unique vals train: {:5d},  n_unique vals test: {:5d}'\n",
    "              .format(feat, X_train[feat].nunique(), X_test[feat].nunique()))\n",
    "        if X_train[feat].nunique() <= 21:\n",
    "            if X_test[feat].nunique() != X_train[feat].nunique():\n",
    "                print('X_test[feat].unique(): {}'.format(X_test[feat].unique()))\n",
    "                print('X_train[feat].unique(): {}'.format(X_train[feat].unique()))\n",
    "#         print('n_unique values in test but not train data: {:5d}'.format(np.intersect1d(X_train[feat].unique(), X_test[feat].unique())))\n",
    "#         print('Feat: {}, number of unique vals: {}, unique vals: {}'.format(feat, X_train[feat].nunique(), X_train[feat].unique()))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We quickly see some interesting things. Some features have thousands of unique values for both the training and testing data. Creating dummy variables for these features would cause the dimensionality of the data to explode, so we'll have to take a closer look at those.\n",
    "\n",
    "Other features have a relatively small number of features (less than ~20 features), which may be reasonable candidates for retaining as categorical features. Creating dummy variables creates a column for each unique value for a given feature, but we see from things like 'scheme_management' or 'extraction_type' that the training and testing data have slightly different numbers of unique values. If the testing data categorical features that aren't in the training data, then the model we trained will have no . \n",
    "\n",
    "\n",
    "\n",
    "One feature has only 1 unique value (recorded_by) in both the training and testing data, and per the check below, the value is the same for both the testing and training data. As all of the values for this feature are the same, it's not providing any new information, so we can drop it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GeoData Consultants Ltd\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(X_train['recorded_by'].iloc[0])\n",
    "X_train['recorded_by'].iloc[0] == X_test['recorded_by'].iloc[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "drop_cols = ['recorded_by']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The set of indices in both the training and testing sets: []\n"
     ]
    }
   ],
   "source": [
    "X_train_inds = X_train.index.values\n",
    "X_test_inds = X_test.index.values\n",
    "print('The set of indices in both the training and testing sets: {}'\n",
    "      .format(np.intersect1d(X_train_inds, X_test_inds)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "# X_train.loc[X_train_inds, :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([], dtype=int64)"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:py36]",
   "language": "python",
   "name": "conda-env-py36-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}