PHL_Property_Assessment.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import re\n",
    "import locale\n",
    "\n",
    "from collections import defaultdict\n",
    "import matplotlib.pyplot as plt; plt.rcdefaults()\n",
    "\n",
    "from sklearn.metrics import accuracy_score, classification_report, confusion_matrix\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.preprocessing import scale\n",
    "from sklearn import tree\n",
    "from sklearn import metrics\n",
    "\n",
    "import statsmodels.api as sm\n",
    "\n",
    "import graphviz\n",
    "import gmaps"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def Diff(li1, li2): \n",
    "    li_dif = [i for i in li1 + li2 if i not in li1 or i not in li2] \n",
    "\n",
    "def nested_dd():\n",
    "    return defaultdict(nested_dd)\n",
    "\n",
    "def Diff(li1, li2): \n",
    "    li_dif = [i for i in li1 + li2 if i not in li1 or i not in li2] \n",
    "    return li_dif\n",
    "\n",
    "def getLatLngByZip(zip,df):\n",
    "    temp=df[df['zip_code'] == zip]\n",
    "    temp=temp[['lat','lng']].drop_duplicates()\n",
    "    return temp.values.tolist()\n",
    "\n",
    "def rmse(predictions,actual):\n",
    "    return np.sqrt(((prediction-actual)**2).mean())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load and View Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "assess_df = pd.read_csv('data/assessments.csv')\n",
    "# Need to set dtype because several columns have mixed data types, setting them all to strings\n",
    "opa_df= pd.read_csv('data/opa_properties_public.csv',dtype={'assessment_date': str,'book_and_page':str, 'cross_reference':str,\n",
    "                                                      'other_building':str, 'sewer':str, 'unfinished':str, 'unit':str, \n",
    "                                                      'utility':str,'year_built':str, 'zip_code':str})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 4280495 entries, 0 to 4280494\n",
      "Data columns (total 7 columns):\n",
      "parcel_number       float64\n",
      "year                int64\n",
      "market_value        float64\n",
      "taxable_land        float64\n",
      "taxable_building    float64\n",
      "exempt_land         float64\n",
      "exempt_building     float64\n",
      "dtypes: float64(6), int64(1)\n",
      "memory usage: 228.6 MB\n",
      "None \n",
      "\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 581298 entries, 0 to 581297\n",
      "Data columns (total 77 columns):\n",
      "number_of_rooms              581287 non-null float64\n",
      "assessment_date              146 non-null object\n",
      "beginning_point              570162 non-null object\n",
      "book_and_page                578466 non-null object\n",
      "building_code                581296 non-null object\n",
      "building_code_description    581296 non-null object\n",
      "category_code                581296 non-null float64\n",
      "category_code_description    581296 non-null object\n",
      "census_tract                 581298 non-null float64\n",
      "central_air                  283461 non-null object\n",
      "cross_reference              120640 non-null object\n",
      "date_exterior_condition      365696 non-null object\n",
      "depth                        581294 non-null float64\n",
      "exempt_building              581238 non-null float64\n",
      "exempt_land                  581238 non-null float64\n",
      "exterior_condition           555381 non-null float64\n",
      "fireplaces                   581276 non-null float64\n",
      "frontage                     581294 non-null float64\n",
      "fuel                         10312 non-null object\n",
      "garage_spaces                581276 non-null float64\n",
      "garage_type                  530759 non-null object\n",
      "general_construction         516864 non-null object\n",
      "geographic_ward              572121 non-null float64\n",
      "homestead_exemption          554408 non-null float64\n",
      "house_extension              581298 non-null int64\n",
      "house_number                 581298 non-null int64\n",
      "interior_condition           554769 non-null float64\n",
      "location                     581298 non-null object\n",
      "mailing_address_1            33101 non-null object\n",
      "mailing_address_2            11362 non-null object\n",
      "mailing_care_of              24070 non-null object\n",
      "mailing_city_state           219966 non-null object\n",
      "mailing_street               219967 non-null object\n",
      "mailing_zip                  219732 non-null object\n",
      "market_value                 581238 non-null float64\n",
      "market_value_date            0 non-null float64\n",
      "number_of_bathrooms          581287 non-null float64\n",
      "number_of_bedrooms           581287 non-null float64\n",
      "basements                    324189 non-null object\n",
      "number_stories               581287 non-null float64\n",
      "off_street_open              581231 non-null float64\n",
      "other_building               1331 non-null object\n",
      "owner_1                      581296 non-null object\n",
      "owner_2                      206750 non-null object\n",
      "parcel_number                581298 non-null int64\n",
      "parcel_shape                 575936 non-null object\n",
      "quality_grade                26424 non-null float64\n",
      "recording_date               581286 non-null object\n",
      "registry_number              580830 non-null object\n",
      "sale_date                    581266 non-null object\n",
      "sale_price                   581298 non-null float64\n",
      "separate_utilities           25253 non-null object\n",
      "sewer                        8117 non-null object\n",
      "site_type                    314957 non-null object\n",
      "state_code                   581264 non-null float64\n",
      "street_code                  581298 non-null int64\n",
      "street_designation           581293 non-null object\n",
      "street_direction             226542 non-null object\n",
      "street_name                  581298 non-null object\n",
      "suffix                       2909 non-null object\n",
      "taxable_building             581238 non-null float64\n",
      "taxable_land                 581238 non-null float64\n",
      "topography                   542406 non-null object\n",
      "total_area                   581294 non-null float64\n",
      "total_livable_area           581287 non-null float64\n",
      "type_heater                  291998 non-null object\n",
      "unfinished                   2948 non-null object\n",
      "unit                         38590 non-null object\n",
      "utility                      9006 non-null object\n",
      "view_type                    560540 non-null object\n",
      "year_built                   581285 non-null object\n",
      "year_built_estimate          440173 non-null object\n",
      "zip_code                     581295 non-null object\n",
      "zoning                       580229 non-null object\n",
      "objectid                     581298 non-null int64\n",
      "lat                          581171 non-null float64\n",
      "lng                          581171 non-null float64\n",
      "dtypes: float64(28), int64(5), object(44)\n",
      "memory usage: 341.5+ MB\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "print(assess_df.info(), \"\\n\")\n",
    "print(opa_df.info())"
   ]
  },
  {
   "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>parcel_number</th>\n",
       "      <th>year</th>\n",
       "      <th>market_value</th>\n",
       "      <th>taxable_land</th>\n",
       "      <th>taxable_building</th>\n",
       "      <th>exempt_land</th>\n",
       "      <th>exempt_building</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2346211</th>\n",
       "      <td>582058600.0</td>\n",
       "      <td>2017</td>\n",
       "      <td>194500.0</td>\n",
       "      <td>58350.0</td>\n",
       "      <td>106150.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>30000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1777317</th>\n",
       "      <td>11440900.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>173600.0</td>\n",
       "      <td>11928.0</td>\n",
       "      <td>161672.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1273005</th>\n",
       "      <td>622389500.0</td>\n",
       "      <td>2014</td>\n",
       "      <td>83600.0</td>\n",
       "      <td>15235.0</td>\n",
       "      <td>68365.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>585677</th>\n",
       "      <td>223225800.0</td>\n",
       "      <td>2017</td>\n",
       "      <td>251400.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>37710.0</td>\n",
       "      <td>213690.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3500233</th>\n",
       "      <td>151306000.0</td>\n",
       "      <td>2016</td>\n",
       "      <td>412000.0</td>\n",
       "      <td>65431.0</td>\n",
       "      <td>346569.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>242660</th>\n",
       "      <td>663088600.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>181000.0</td>\n",
       "      <td>48906.0</td>\n",
       "      <td>132094.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4076023</th>\n",
       "      <td>583010315.0</td>\n",
       "      <td>2019</td>\n",
       "      <td>327000.0</td>\n",
       "      <td>93195.0</td>\n",
       "      <td>193805.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>40000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3901504</th>\n",
       "      <td>364028200.0</td>\n",
       "      <td>2019</td>\n",
       "      <td>77100.0</td>\n",
       "      <td>15100.0</td>\n",
       "      <td>62000.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1713769</th>\n",
       "      <td>581095800.0</td>\n",
       "      <td>2014</td>\n",
       "      <td>324000.0</td>\n",
       "      <td>281130.0</td>\n",
       "      <td>42870.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1685392</th>\n",
       "      <td>582397400.0</td>\n",
       "      <td>2020</td>\n",
       "      <td>260600.0</td>\n",
       "      <td>78180.0</td>\n",
       "      <td>182420.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         parcel_number  year  market_value  taxable_land  taxable_building  \\\n",
       "2346211    582058600.0  2017      194500.0       58350.0          106150.0   \n",
       "1777317     11440900.0  2015      173600.0       11928.0          161672.0   \n",
       "1273005    622389500.0  2014       83600.0       15235.0           68365.0   \n",
       "585677     223225800.0  2017      251400.0           0.0               0.0   \n",
       "3500233    151306000.0  2016      412000.0       65431.0          346569.0   \n",
       "242660     663088600.0  2015      181000.0       48906.0          132094.0   \n",
       "4076023    583010315.0  2019      327000.0       93195.0          193805.0   \n",
       "3901504    364028200.0  2019       77100.0       15100.0           62000.0   \n",
       "1713769    581095800.0  2014      324000.0      281130.0           42870.0   \n",
       "1685392    582397400.0  2020      260600.0       78180.0          182420.0   \n",
       "\n",
       "         exempt_land  exempt_building  \n",
       "2346211          0.0          30000.0  \n",
       "1777317          0.0              0.0  \n",
       "1273005          0.0              0.0  \n",
       "585677       37710.0         213690.0  \n",
       "3500233          0.0              0.0  \n",
       "242660           0.0              0.0  \n",
       "4076023          0.0          40000.0  \n",
       "3901504          0.0              0.0  \n",
       "1713769          0.0              0.0  \n",
       "1685392          0.0              0.0  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "assess_df.sample(n=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>number_of_rooms</th>\n",
       "      <th>assessment_date</th>\n",
       "      <th>beginning_point</th>\n",
       "      <th>book_and_page</th>\n",
       "      <th>building_code</th>\n",
       "      <th>building_code_description</th>\n",
       "      <th>category_code</th>\n",
       "      <th>category_code_description</th>\n",
       "      <th>census_tract</th>\n",
       "      <th>central_air</th>\n",
       "      <th>...</th>\n",
       "      <th>unit</th>\n",
       "      <th>utility</th>\n",
       "      <th>view_type</th>\n",
       "      <th>year_built</th>\n",
       "      <th>year_built_estimate</th>\n",
       "      <th>zip_code</th>\n",
       "      <th>zoning</th>\n",
       "      <th>objectid</th>\n",
       "      <th>lat</th>\n",
       "      <th>lng</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>511634</th>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19' N OXFORD ST</td>\n",
       "      <td>0157496</td>\n",
       "      <td>U50</td>\n",
       "      <td>ROW CONV/APT 3 STY MASON</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Multi Family</td>\n",
       "      <td>731.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I</td>\n",
       "      <td>1915</td>\n",
       "      <td>Y</td>\n",
       "      <td>191213202</td>\n",
       "      <td>RM1</td>\n",
       "      <td>498324689</td>\n",
       "      <td>39.977860</td>\n",
       "      <td>-75.161655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>383191</th>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>100' E LORETTO AVE</td>\n",
       "      <td>0623936</td>\n",
       "      <td>B10</td>\n",
       "      <td>DET W/GAR 1 STY MASONRY</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Single Family</td>\n",
       "      <td>720.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I</td>\n",
       "      <td>1940</td>\n",
       "      <td>NaN</td>\n",
       "      <td>191522614</td>\n",
       "      <td>RSA3</td>\n",
       "      <td>498195034</td>\n",
       "      <td>40.065132</td>\n",
       "      <td>-75.054700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>186805</th>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>107' N OF COLUMBIA</td>\n",
       "      <td>0000000</td>\n",
       "      <td>O30</td>\n",
       "      <td>ROW 2 STY MASONRY</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Single Family</td>\n",
       "      <td>930.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I</td>\n",
       "      <td>1925</td>\n",
       "      <td>NaN</td>\n",
       "      <td>191212811</td>\n",
       "      <td>RSA5</td>\n",
       "      <td>498004692</td>\n",
       "      <td>39.981675</td>\n",
       "      <td>-75.177893</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>136101</th>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>114' 1/2\"SE HORROCK</td>\n",
       "      <td>2973791</td>\n",
       "      <td>Q30</td>\n",
       "      <td>ROW W/DET GAR 2 STY MAS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Single Family</td>\n",
       "      <td>130.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I</td>\n",
       "      <td>1945</td>\n",
       "      <td>Y</td>\n",
       "      <td>191243136</td>\n",
       "      <td>RM1</td>\n",
       "      <td>497954614</td>\n",
       "      <td>40.019070</td>\n",
       "      <td>-75.089720</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7179</th>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>298' W OF 9TH ST</td>\n",
       "      <td>2852457</td>\n",
       "      <td>O30</td>\n",
       "      <td>ROW 2 STY MASONRY</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Single Family</td>\n",
       "      <td>881.0</td>\n",
       "      <td>Y</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I</td>\n",
       "      <td>1920</td>\n",
       "      <td>NaN</td>\n",
       "      <td>191481112</td>\n",
       "      <td>RSA5</td>\n",
       "      <td>497817523</td>\n",
       "      <td>39.928163</td>\n",
       "      <td>-75.161406</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>251760</th>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>59' S WOLF</td>\n",
       "      <td>1069235</td>\n",
       "      <td>O30</td>\n",
       "      <td>ROW 2 STY MASONRY</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Single Family</td>\n",
       "      <td>160.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I</td>\n",
       "      <td>1940</td>\n",
       "      <td>Y</td>\n",
       "      <td>191483702</td>\n",
       "      <td>RSA5</td>\n",
       "      <td>498065346</td>\n",
       "      <td>39.920456</td>\n",
       "      <td>-75.161201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>301484</th>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>208'1 1/8\" N.</td>\n",
       "      <td>1644816</td>\n",
       "      <td>O30</td>\n",
       "      <td>ROW 2 STY MASONRY</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Single Family</td>\n",
       "      <td>410.0</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I</td>\n",
       "      <td>1942</td>\n",
       "      <td>Y</td>\n",
       "      <td>191204529</td>\n",
       "      <td>RSA5</td>\n",
       "      <td>498112758</td>\n",
       "      <td>40.018136</td>\n",
       "      <td>-75.117477</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>361267</th>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>320'8 1/2\" E OF</td>\n",
       "      <td>0063031</td>\n",
       "      <td>K30</td>\n",
       "      <td>S/D W/B GAR 2 STY MASONRY</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Single Family</td>\n",
       "      <td>131.0</td>\n",
       "      <td>Y</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I</td>\n",
       "      <td>1950</td>\n",
       "      <td>Y</td>\n",
       "      <td>191500000</td>\n",
       "      <td>RSA3</td>\n",
       "      <td>498170460</td>\n",
       "      <td>40.074023</td>\n",
       "      <td>-75.175792</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>315147</th>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>122  1\" N VENANGO ST</td>\n",
       "      <td>0463313</td>\n",
       "      <td>O30</td>\n",
       "      <td>ROW 2 STY MASONRY</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Single Family</td>\n",
       "      <td>610.0</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I</td>\n",
       "      <td>1946</td>\n",
       "      <td>Y</td>\n",
       "      <td>191345519</td>\n",
       "      <td>RSA5</td>\n",
       "      <td>498120333</td>\n",
       "      <td>39.991527</td>\n",
       "      <td>-75.094660</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>538339</th>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>SEC CRANSTON RD</td>\n",
       "      <td>1695976</td>\n",
       "      <td>590</td>\n",
       "      <td>RES CONDO 5+ STY MASONRY</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Single Family</td>\n",
       "      <td>230.0</td>\n",
       "      <td>Y</td>\n",
       "      <td>...</td>\n",
       "      <td>000017O</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I</td>\n",
       "      <td>1960</td>\n",
       "      <td>Y</td>\n",
       "      <td>191312054</td>\n",
       "      <td>RM1</td>\n",
       "      <td>498350370</td>\n",
       "      <td>39.998782</td>\n",
       "      <td>-75.208691</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10 rows × 77 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        number_of_rooms assessment_date            beginning_point  \\\n",
       "511634              0.0             NaN  19' N OXFORD ST             \n",
       "383191              0.0             NaN  100' E LORETTO AVE          \n",
       "186805              4.0             NaN  107' N OF COLUMBIA          \n",
       "136101              6.0             NaN  114' 1/2\"SE HORROCK         \n",
       "7179                6.0             NaN  298' W OF 9TH ST            \n",
       "251760              6.0             NaN  59' S WOLF                  \n",
       "301484              6.0             NaN  208'1 1/8\" N.               \n",
       "361267              6.0             NaN  320'8 1/2\" E OF             \n",
       "315147              4.0             NaN  122  1\" N VENANGO ST        \n",
       "538339              0.0             NaN  SEC CRANSTON RD             \n",
       "\n",
       "       book_and_page building_code  building_code_description  category_code  \\\n",
       "511634       0157496         U50     ROW CONV/APT 3 STY MASON            2.0   \n",
       "383191       0623936         B10      DET W/GAR 1 STY MASONRY            1.0   \n",
       "186805       0000000         O30            ROW 2 STY MASONRY            1.0   \n",
       "136101       2973791         Q30      ROW W/DET GAR 2 STY MAS            1.0   \n",
       "7179         2852457         O30            ROW 2 STY MASONRY            1.0   \n",
       "251760       1069235         O30            ROW 2 STY MASONRY            1.0   \n",
       "301484       1644816         O30            ROW 2 STY MASONRY            1.0   \n",
       "361267       0063031         K30    S/D W/B GAR 2 STY MASONRY            1.0   \n",
       "315147       0463313         O30            ROW 2 STY MASONRY            1.0   \n",
       "538339       1695976         590     RES CONDO 5+ STY MASONRY            1.0   \n",
       "\n",
       "       category_code_description  census_tract central_air  ...     unit  \\\n",
       "511634              Multi Family         731.0         NaN  ...      NaN   \n",
       "383191             Single Family         720.0         NaN  ...      NaN   \n",
       "186805             Single Family         930.0         NaN  ...      NaN   \n",
       "136101             Single Family         130.0         NaN  ...      NaN   \n",
       "7179               Single Family         881.0           Y  ...      NaN   \n",
       "251760             Single Family         160.0         NaN  ...      NaN   \n",
       "301484             Single Family         410.0           N  ...      NaN   \n",
       "361267             Single Family         131.0           Y  ...      NaN   \n",
       "315147             Single Family         610.0           N  ...      NaN   \n",
       "538339             Single Family         230.0           Y  ...  000017O   \n",
       "\n",
       "       utility  view_type  year_built  year_built_estimate   zip_code  zoning  \\\n",
       "511634     NaN          I        1915                    Y  191213202   RM1     \n",
       "383191     NaN          I        1940                  NaN  191522614   RSA3    \n",
       "186805     NaN          I        1925                  NaN  191212811   RSA5    \n",
       "136101     NaN          I        1945                    Y  191243136   RM1     \n",
       "7179       NaN          I        1920                  NaN  191481112   RSA5    \n",
       "251760     NaN          I        1940                    Y  191483702   RSA5    \n",
       "301484     NaN          I        1942                    Y  191204529   RSA5    \n",
       "361267     NaN          I        1950                    Y  191500000   RSA3    \n",
       "315147     NaN          I        1946                    Y  191345519   RSA5    \n",
       "538339     NaN          I        1960                    Y  191312054   RM1     \n",
       "\n",
       "         objectid        lat        lng  \n",
       "511634  498324689  39.977860 -75.161655  \n",
       "383191  498195034  40.065132 -75.054700  \n",
       "186805  498004692  39.981675 -75.177893  \n",
       "136101  497954614  40.019070 -75.089720  \n",
       "7179    497817523  39.928163 -75.161406  \n",
       "251760  498065346  39.920456 -75.161201  \n",
       "301484  498112758  40.018136 -75.117477  \n",
       "361267  498170460  40.074023 -75.175792  \n",
       "315147  498120333  39.991527 -75.094660  \n",
       "538339  498350370  39.998782 -75.208691  \n",
       "\n",
       "[10 rows x 77 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opa_df.sample(n=10)"
   ]
  },
  {
   "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>number_of_rooms</th>\n",
       "      <th>category_code</th>\n",
       "      <th>census_tract</th>\n",
       "      <th>depth</th>\n",
       "      <th>exempt_building</th>\n",
       "      <th>exempt_land</th>\n",
       "      <th>exterior_condition</th>\n",
       "      <th>fireplaces</th>\n",
       "      <th>frontage</th>\n",
       "      <th>garage_spaces</th>\n",
       "      <th>...</th>\n",
       "      <th>sale_price</th>\n",
       "      <th>state_code</th>\n",
       "      <th>street_code</th>\n",
       "      <th>taxable_building</th>\n",
       "      <th>taxable_land</th>\n",
       "      <th>total_area</th>\n",
       "      <th>total_livable_area</th>\n",
       "      <th>objectid</th>\n",
       "      <th>lat</th>\n",
       "      <th>lng</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>581287.000000</td>\n",
       "      <td>581296.000000</td>\n",
       "      <td>581298.000000</td>\n",
       "      <td>5.812940e+05</td>\n",
       "      <td>5.812380e+05</td>\n",
       "      <td>5.812380e+05</td>\n",
       "      <td>555381.000000</td>\n",
       "      <td>581276.000000</td>\n",
       "      <td>5.812940e+05</td>\n",
       "      <td>581276.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>5.812980e+05</td>\n",
       "      <td>581264.000000</td>\n",
       "      <td>581298.000000</td>\n",
       "      <td>5.812380e+05</td>\n",
       "      <td>5.812380e+05</td>\n",
       "      <td>5.812940e+05</td>\n",
       "      <td>5.812870e+05</td>\n",
       "      <td>5.812980e+05</td>\n",
       "      <td>581171.000000</td>\n",
       "      <td>581171.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>3.765641</td>\n",
       "      <td>1.609507</td>\n",
       "      <td>463.700641</td>\n",
       "      <td>1.108392e+02</td>\n",
       "      <td>7.476410e+04</td>\n",
       "      <td>1.629839e+04</td>\n",
       "      <td>3.740045</td>\n",
       "      <td>0.031802</td>\n",
       "      <td>1.098852e+03</td>\n",
       "      <td>0.349196</td>\n",
       "      <td>...</td>\n",
       "      <td>2.969122e+05</td>\n",
       "      <td>1514.381453</td>\n",
       "      <td>54188.622355</td>\n",
       "      <td>1.471248e+05</td>\n",
       "      <td>5.279799e+04</td>\n",
       "      <td>5.039855e+03</td>\n",
       "      <td>2.250229e+03</td>\n",
       "      <td>4.980996e+08</td>\n",
       "      <td>39.999830</td>\n",
       "      <td>-75.143500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>3.105311</td>\n",
       "      <td>1.435636</td>\n",
       "      <td>288.771231</td>\n",
       "      <td>4.269818e+03</td>\n",
       "      <td>1.808741e+06</td>\n",
       "      <td>8.067065e+05</td>\n",
       "      <td>1.015864</td>\n",
       "      <td>0.235121</td>\n",
       "      <td>5.842447e+05</td>\n",
       "      <td>0.934259</td>\n",
       "      <td>...</td>\n",
       "      <td>1.121295e+07</td>\n",
       "      <td>1341.414954</td>\n",
       "      <td>24564.201119</td>\n",
       "      <td>1.608646e+06</td>\n",
       "      <td>4.269669e+05</td>\n",
       "      <td>2.965323e+05</td>\n",
       "      <td>1.649796e+04</td>\n",
       "      <td>1.678064e+05</td>\n",
       "      <td>0.050866</td>\n",
       "      <td>0.065903</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1001.000000</td>\n",
       "      <td>11020.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>4.978088e+08</td>\n",
       "      <td>39.875131</td>\n",
       "      <td>-75.274389</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>212.000000</td>\n",
       "      <td>5.700000e+01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.500000e+01</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1001.000000</td>\n",
       "      <td>33120.000000</td>\n",
       "      <td>2.999100e+04</td>\n",
       "      <td>1.122000e+04</td>\n",
       "      <td>8.823700e+02</td>\n",
       "      <td>1.024000e+03</td>\n",
       "      <td>4.979543e+08</td>\n",
       "      <td>39.957732</td>\n",
       "      <td>-75.182658</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>6.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>440.000000</td>\n",
       "      <td>7.700000e+01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.600000e+01</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>3.800000e+04</td>\n",
       "      <td>1001.000000</td>\n",
       "      <td>54840.000000</td>\n",
       "      <td>7.633000e+04</td>\n",
       "      <td>2.175650e+04</td>\n",
       "      <td>1.280000e+03</td>\n",
       "      <td>1.230000e+03</td>\n",
       "      <td>4.980996e+08</td>\n",
       "      <td>39.996515</td>\n",
       "      <td>-75.155170</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>6.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>722.000000</td>\n",
       "      <td>1.000000e+02</td>\n",
       "      <td>4.000000e+04</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.041000e+01</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.285000e+05</td>\n",
       "      <td>1001.000000</td>\n",
       "      <td>76760.000000</td>\n",
       "      <td>1.355600e+05</td>\n",
       "      <td>4.560000e+04</td>\n",
       "      <td>2.050410e+03</td>\n",
       "      <td>1.586000e+03</td>\n",
       "      <td>4.982449e+08</td>\n",
       "      <td>40.039884</td>\n",
       "      <td>-75.104737</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>89.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>982.000000</td>\n",
       "      <td>1.829520e+06</td>\n",
       "      <td>4.304176e+08</td>\n",
       "      <td>3.418426e+08</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>4.320260e+08</td>\n",
       "      <td>97.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>9.487291e+08</td>\n",
       "      <td>9904.000000</td>\n",
       "      <td>89790.000000</td>\n",
       "      <td>3.048985e+08</td>\n",
       "      <td>7.411128e+07</td>\n",
       "      <td>2.076941e+08</td>\n",
       "      <td>2.500000e+06</td>\n",
       "      <td>4.983903e+08</td>\n",
       "      <td>40.137705</td>\n",
       "      <td>-74.958190</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 33 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       number_of_rooms  category_code   census_tract         depth  \\\n",
       "count    581287.000000  581296.000000  581298.000000  5.812940e+05   \n",
       "mean          3.765641       1.609507     463.700641  1.108392e+02   \n",
       "std           3.105311       1.435636     288.771231  4.269818e+03   \n",
       "min           0.000000       1.000000       0.000000  0.000000e+00   \n",
       "25%           0.000000       1.000000     212.000000  5.700000e+01   \n",
       "50%           6.000000       1.000000     440.000000  7.700000e+01   \n",
       "75%           6.000000       1.000000     722.000000  1.000000e+02   \n",
       "max          89.000000       6.000000     982.000000  1.829520e+06   \n",
       "\n",
       "       exempt_building   exempt_land  exterior_condition     fireplaces  \\\n",
       "count     5.812380e+05  5.812380e+05       555381.000000  581276.000000   \n",
       "mean      7.476410e+04  1.629839e+04            3.740045       0.031802   \n",
       "std       1.808741e+06  8.067065e+05            1.015864       0.235121   \n",
       "min       0.000000e+00  0.000000e+00            0.000000       0.000000   \n",
       "25%       0.000000e+00  0.000000e+00            4.000000       0.000000   \n",
       "50%       0.000000e+00  0.000000e+00            4.000000       0.000000   \n",
       "75%       4.000000e+04  0.000000e+00            4.000000       0.000000   \n",
       "max       4.304176e+08  3.418426e+08            7.000000      10.000000   \n",
       "\n",
       "           frontage  garage_spaces  ...    sale_price     state_code  \\\n",
       "count  5.812940e+05  581276.000000  ...  5.812980e+05  581264.000000   \n",
       "mean   1.098852e+03       0.349196  ...  2.969122e+05    1514.381453   \n",
       "std    5.842447e+05       0.934259  ...  1.121295e+07    1341.414954   \n",
       "min    0.000000e+00       0.000000  ...  0.000000e+00    1001.000000   \n",
       "25%    1.500000e+01       0.000000  ...  1.000000e+00    1001.000000   \n",
       "50%    1.600000e+01       0.000000  ...  3.800000e+04    1001.000000   \n",
       "75%    2.041000e+01       1.000000  ...  1.285000e+05    1001.000000   \n",
       "max    4.320260e+08      97.000000  ...  9.487291e+08    9904.000000   \n",
       "\n",
       "         street_code  taxable_building  taxable_land    total_area  \\\n",
       "count  581298.000000      5.812380e+05  5.812380e+05  5.812940e+05   \n",
       "mean    54188.622355      1.471248e+05  5.279799e+04  5.039855e+03   \n",
       "std     24564.201119      1.608646e+06  4.269669e+05  2.965323e+05   \n",
       "min     11020.000000      0.000000e+00  0.000000e+00  0.000000e+00   \n",
       "25%     33120.000000      2.999100e+04  1.122000e+04  8.823700e+02   \n",
       "50%     54840.000000      7.633000e+04  2.175650e+04  1.280000e+03   \n",
       "75%     76760.000000      1.355600e+05  4.560000e+04  2.050410e+03   \n",
       "max     89790.000000      3.048985e+08  7.411128e+07  2.076941e+08   \n",
       "\n",
       "       total_livable_area      objectid            lat            lng  \n",
       "count        5.812870e+05  5.812980e+05  581171.000000  581171.000000  \n",
       "mean         2.250229e+03  4.980996e+08      39.999830     -75.143500  \n",
       "std          1.649796e+04  1.678064e+05       0.050866       0.065903  \n",
       "min          0.000000e+00  4.978088e+08      39.875131     -75.274389  \n",
       "25%          1.024000e+03  4.979543e+08      39.957732     -75.182658  \n",
       "50%          1.230000e+03  4.980996e+08      39.996515     -75.155170  \n",
       "75%          1.586000e+03  4.982449e+08      40.039884     -75.104737  \n",
       "max          2.500000e+06  4.983903e+08      40.137705     -74.958190  \n",
       "\n",
       "[8 rows x 33 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opa_df.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Transform DataFrames"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "parcel_number       207903\n",
       "year                     0\n",
       "market_value             0\n",
       "taxable_land             0\n",
       "taxable_building         0\n",
       "exempt_land              0\n",
       "exempt_building          0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#count null values\n",
    "assess_df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Removing any rows that have NULL values in the parcel_number column\n",
    "assess_df=assess_df[~(assess_df['parcel_number'].isnull())]\n",
    "opa_df=opa_df[~(opa_df['parcel_number'].isnull())]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "parcel_number       0\n",
      "year                0\n",
      "market_value        0\n",
      "taxable_land        0\n",
      "taxable_building    0\n",
      "exempt_land         0\n",
      "exempt_building     0\n",
      "dtype: int64 \n",
      "\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "print(assess_df.isnull().sum(),\"\\n\")\n",
    "print(opa_df['parcel_number'].isnull().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Add exempt_total column\n",
    "assess_df['exempt_total'] = assess_df['exempt_building'] + assess_df['exempt_land']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "assess_df_merged = pd.merge(assess_df, opa_df[['building_code','building_code_description','category_code','category_code_description',\n",
    "                                               'year_built','lat','lng','homestead_exemption','number_of_bathrooms','number_of_bedrooms',\n",
    "                                               'owner_1','owner_2','sale_date','sale_price','total_livable_area','zoning', 'parcel_number']]\n",
    "                            ,on = 'parcel_number')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Taking columns needed from assessments.csv\n",
    "df1 = assess_df[['parcel_number','year','exempt_total']]\n",
    "# Taking columns needed from opa_properties_public.csv\n",
    "df2 = opa_df[['parcel_number','lat','lng','zip_code']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Merge dataframes into 1\n",
    "df=pd.merge(df1, df2, on='parcel_number')\n",
    "# Removing any rows that have NaN values in the lat or lng column\n",
    "df=df[~(df['lat'].isnull())]\n",
    "df=df[~(df['lng'].isnull())]\n",
    "# just get first 5 of zipcode\n",
    "df['zip_code']=df['zip_code'].str.extract('([0-9]{5})', expand=True)\n",
    "df['zip_code'].fillna('1', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating 2 dataframes, one that has the 10 year abatements, and one that does not\n",
    "#40k is the homestead abatement\n",
    "df_a=df[df['exempt_total'] > 40000]\n",
    "df_n=df[df['exempt_total'] <= 40000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "#create bool mask for different abatements\n",
    "assess_df_merged['homestead_exemption'] = ((assess_df_merged['category_code'] == 1) & \n",
    "                                    ((assess_df_merged['year'] <= 2018) & (assess_df_merged['exempt_building'] == 30000) |\n",
    "                                    (assess_df_merged['year'] >= 2019) & (assess_df_merged['exempt_building'] == 40000)))\n",
    "assess_df_merged['abatement10'] = ((assess_df_merged['year'] <= 2018) & (assess_df_merged['exempt_building'] > 30000) |\n",
    "                                    (assess_df_merged['year'] >= 2019) & (assess_df_merged['exempt_building'] > 40000))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Group By"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "grouped_property = assess_df_merged.groupby(['year'])\n",
    "\n",
    "#Average assessed property value per year\n",
    "avg_mv = grouped_property.mean()['market_value']\n",
    "#Median assessed property value per year\n",
    "med_mv = grouped_property.median()['market_value']\n",
    "\n",
    "#Percentage of homestead and 10 year abated properties\n",
    "hm = grouped_property.mean()['homestead_exemption']\n",
    "ab = grouped_property.mean()['abatement10']\n",
    "\n",
    "#count of properties per year\n",
    "count = grouped_property.count()['parcel_number']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "years = avg_mv.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "## set the figure dimensions\n",
    "line_fig = plt.figure(figsize = (10, 6))\n",
    "    \n",
    "## make a line plot\n",
    "_  = plt.plot(years, avg_mv, label = \"Mean\")\n",
    "_  = plt.plot(years, med_mv, label = \"Median\")\n",
    "\n",
    "## Set the tick and label fontsize\n",
    "plt.tick_params(labelsize = 15)\n",
    "\n",
    "## Set the title\n",
    "_ = plt.title(\"Median and Mean Property Market Value\", fontsize = 15)\n",
    "\n",
    "## Set the y-label\n",
    "_ = plt.ylabel(\"Market Value ($)\", fontsize = 15)\n",
    "\n",
    "## Set the x-label\n",
    "_ = plt.xlabel(\"Year\", fontsize = 15)\n",
    "\n",
    "## Set Legend\n",
    "_ = plt.legend(loc='upper left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Calculate optimal width\n",
    "width = np.min(np.diff(years))/3\n",
    "\n",
    "## set the figure dimensions\n",
    "bar_fig = plt.figure(figsize = (10, 6))\n",
    "ax = bar_fig.add_subplot(111)\n",
    "    \n",
    "## make a line plot\n",
    "_ = ax.bar(years-width, hm*count, width, label = \"Home Exemptions\")\n",
    "_ = ax.bar(years, ab*count, width, label = \"10 Year Abatements\")\n",
    "_ = ax.set_yticklabels(['{:,}'.format(int(x)) for x in ax.get_yticks().tolist()])\n",
    "\n",
    "## Set the tick and label fontsize\n",
    "plt.tick_params(labelsize = 15)\n",
    "\n",
    "## Set the title\n",
    "_ = plt.title(\"Properties with Preferential Taxes by Year\", fontsize = 15)\n",
    "\n",
    "## Set the y-label\n",
    "_ = plt.ylabel(\"Count\", fontsize = 15)\n",
    "\n",
    "## Set the x-label\n",
    "_ = plt.xlabel(\"Year\", fontsize = 15)\n",
    "\n",
    "## Set Legend\n",
    "_ = plt.legend(loc='upper left')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The Homestead exemption was passed by City Council in 2013 with implementation beginning in 2014 at $30,000. Data is missing for these properties in 2014 and 2015."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "grouped_property_ccd = assess_df_merged.groupby(['category_code_description'])\n",
    "count_ccd = grouped_property_ccd.count()['abatement10']\n",
    "mean_ccd = grouped_property_ccd.mean()['abatement10']"
   ]
  },
  {
   "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>COUNT PROPERTY</th>\n",
       "      <th>COUNT ABATED PROPERTY</th>\n",
       "      <th>PERCENTAGE</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_code_description</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Industrial</th>\n",
       "      <td>30835</td>\n",
       "      <td>1626.0</td>\n",
       "      <td>5.273228</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Commercial</th>\n",
       "      <td>97764</td>\n",
       "      <td>27657.0</td>\n",
       "      <td>28.289554</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mixed Use</th>\n",
       "      <td>100854</td>\n",
       "      <td>5560.0</td>\n",
       "      <td>5.512920</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Multi Family</th>\n",
       "      <td>295096</td>\n",
       "      <td>23360.0</td>\n",
       "      <td>7.916068</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Vacant Land</th>\n",
       "      <td>309393</td>\n",
       "      <td>978.0</td>\n",
       "      <td>0.316103</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Single Family</th>\n",
       "      <td>3220669</td>\n",
       "      <td>218381.0</td>\n",
       "      <td>6.780610</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           COUNT PROPERTY  COUNT ABATED PROPERTY  PERCENTAGE\n",
       "category_code_description                                                   \n",
       "Industrial                          30835                 1626.0    5.273228\n",
       "Commercial                          97764                27657.0   28.289554\n",
       "Mixed Use                          100854                 5560.0    5.512920\n",
       "Multi Family                       295096                23360.0    7.916068\n",
       "Vacant Land                        309393                  978.0    0.316103\n",
       "Single Family                     3220669               218381.0    6.780610"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ccd_table = pd.DataFrame({\"COUNT PROPERTY\":count_ccd,\n",
    "                   \"COUNT ABATED PROPERTY\":mean_ccd*count_ccd})\n",
    "ccd_table['PERCENTAGE'] = (ccd_table['COUNT ABATED PROPERTY']/ccd_table['COUNT PROPERTY'])*100\n",
    "ccd_table = ccd_table.sort_values(by='COUNT PROPERTY', ascending = True)\n",
    "ccd_table"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are 978 properties with a `category_code_description` of **Vacant Land** with a 10 year tax abatement. This is interesting because abatements are applicable to building improvements and not empty lots. These properties abatements should be reviewed by city officials, or this may be a case of bad data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Example of Abated Home"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>parcel_number</th>\n",
       "      <th>year</th>\n",
       "      <th>market_value</th>\n",
       "      <th>taxable_land</th>\n",
       "      <th>taxable_building</th>\n",
       "      <th>exempt_land</th>\n",
       "      <th>exempt_building</th>\n",
       "      <th>exempt_total</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1110360</th>\n",
       "      <td>888154334.0</td>\n",
       "      <td>2014</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1110359</th>\n",
       "      <td>888154334.0</td>\n",
       "      <td>2015</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1110358</th>\n",
       "      <td>888154334.0</td>\n",
       "      <td>2016</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1110357</th>\n",
       "      <td>888154334.0</td>\n",
       "      <td>2017</td>\n",
       "      <td>265500.0</td>\n",
       "      <td>39825.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>225675.0</td>\n",
       "      <td>225675.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1110356</th>\n",
       "      <td>888154334.0</td>\n",
       "      <td>2018</td>\n",
       "      <td>265500.0</td>\n",
       "      <td>39825.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>225675.0</td>\n",
       "      <td>225675.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4228933</th>\n",
       "      <td>888154334.0</td>\n",
       "      <td>2019</td>\n",
       "      <td>265500.0</td>\n",
       "      <td>39825.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>225675.0</td>\n",
       "      <td>225675.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2252993</th>\n",
       "      <td>888154334.0</td>\n",
       "      <td>2020</td>\n",
       "      <td>265500.0</td>\n",
       "      <td>39825.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>225675.0</td>\n",
       "      <td>225675.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         parcel_number  year  market_value  taxable_land  taxable_building  \\\n",
       "1110360    888154334.0  2014           0.0           0.0               0.0   \n",
       "1110359    888154334.0  2015           0.0           0.0               0.0   \n",
       "1110358    888154334.0  2016           0.0           0.0               0.0   \n",
       "1110357    888154334.0  2017      265500.0       39825.0               0.0   \n",
       "1110356    888154334.0  2018      265500.0       39825.0               0.0   \n",
       "4228933    888154334.0  2019      265500.0       39825.0               0.0   \n",
       "2252993    888154334.0  2020      265500.0       39825.0               0.0   \n",
       "\n",
       "         exempt_land  exempt_building  exempt_total  \n",
       "1110360          0.0              0.0           0.0  \n",
       "1110359          0.0              0.0           0.0  \n",
       "1110358          0.0              0.0           0.0  \n",
       "1110357          0.0         225675.0      225675.0  \n",
       "1110356          0.0         225675.0      225675.0  \n",
       "4228933          0.0         225675.0      225675.0  \n",
       "2252993          0.0         225675.0      225675.0  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Aviad's home\n",
    "#example of a property with a 10 year tax abatement\n",
    "assess_df.loc[assess_df['parcel_number'] == 888154334].sort_values('year')"
   ]
  },
  {
   "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>number_of_rooms</th>\n",
       "      <th>assessment_date</th>\n",
       "      <th>beginning_point</th>\n",
       "      <th>book_and_page</th>\n",
       "      <th>building_code</th>\n",
       "      <th>building_code_description</th>\n",
       "      <th>category_code</th>\n",
       "      <th>category_code_description</th>\n",
       "      <th>census_tract</th>\n",
       "      <th>central_air</th>\n",
       "      <th>...</th>\n",
       "      <th>unit</th>\n",
       "      <th>utility</th>\n",
       "      <th>view_type</th>\n",
       "      <th>year_built</th>\n",
       "      <th>year_built_estimate</th>\n",
       "      <th>zip_code</th>\n",
       "      <th>zoning</th>\n",
       "      <th>objectid</th>\n",
       "      <th>lat</th>\n",
       "      <th>lng</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>537764</th>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>SEC OF 18TH ST</td>\n",
       "      <td>3198301</td>\n",
       "      <td>571</td>\n",
       "      <td>RES CONDO 4 STY MAS+OTH</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Single Family</td>\n",
       "      <td>520.0</td>\n",
       "      <td>Y</td>\n",
       "      <td>...</td>\n",
       "      <td>000000B</td>\n",
       "      <td>A</td>\n",
       "      <td>I</td>\n",
       "      <td>2015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>191302816</td>\n",
       "      <td>RM1</td>\n",
       "      <td>498351383</td>\n",
       "      <td>39.967815</td>\n",
       "      <td>-75.16665</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1 rows × 77 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        number_of_rooms assessment_date            beginning_point  \\\n",
       "537764              0.0             NaN  SEC OF 18TH ST              \n",
       "\n",
       "       book_and_page building_code building_code_description  category_code  \\\n",
       "537764       3198301         571     RES CONDO 4 STY MAS+OTH            1.0   \n",
       "\n",
       "       category_code_description  census_tract central_air  ...     unit  \\\n",
       "537764             Single Family         520.0           Y  ...  000000B   \n",
       "\n",
       "       utility  view_type  year_built  year_built_estimate   zip_code  zoning  \\\n",
       "537764       A          I        2015                  NaN  191302816   RM1     \n",
       "\n",
       "         objectid        lat       lng  \n",
       "537764  498351383  39.967815 -75.16665  \n",
       "\n",
       "[1 rows x 77 columns]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Aviad's home\n",
    "#example of a property with a 10 year tax abatement\n",
    "opa_df.loc[opa_df['parcel_number'] == 888154334]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Map abated homes on a map"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('apikey.txt') as f:\n",
    "    gmap_api_key = f.readline()\n",
    "    f.close\n",
    "    \n",
    "gmaps.configure(api_key = gmap_api_key)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Heatmap of homestead exemptions\n",
    "\n",
    "All Philadelphian property owners who use their property as their primary residence are eligable for the Homestead Exemption. Hence we see high density of properties with this exemption throughout the city."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![Homestead Abatements 1](maps/Heat_Maps/Heat_Map_Homestead_1.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![Homestead Abatements 1](maps/Heat_Maps/Heat_Map_Homestead_2.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "#homestead exemption in 2020 is $45k, however data shows $40k as this was the original proposed amount. \n",
    "#City counsel increased amount after data was published. \n",
    "#Confrimed with Sharon from revenue department that the data is stale.\n",
    "\n",
    "home_expt = opa_df[['homestead_exemption','lat','lng']].loc[opa_df['homestead_exemption'] == 40000]\n",
    "home_expt['lng'] = home_expt['lng'].round(4)\n",
    "home_expt['lat'] = home_expt['lat'].round(4)\n",
    "grouped_homestead = home_expt.groupby(['lat','lng'])\n",
    "a = grouped_homestead.count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "locations = a.index.values\n",
    "weights = a['homestead_exemption'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9d33c044b9624277a509da6e3c8bed5f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Figure(layout=FigureLayout(height='420px'))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Centered on City Hall\n",
    "fig = gmaps.figure(center = (39.9524, -75.1636), zoom_level=11, map_type = 'HYBRID')\n",
    "fig.add_layer(gmaps.heatmap_layer(locations, weights = weights, max_intensity = 100, point_radius = 5))\n",
    "fig"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Heatmap of 10 year tax abatement\n",
    "\n",
    "The ten year tax abatement \"encourages new construction or rehabilitation of properties with the exemption of all or part of the value of the improvement for a set number of years. Abatements help revitalize communities, retain residents, attract home- and business-owners to the city of Philadelphia, and reduce development costs for commercial and residential projects.\" The maps below show high concentration of abated properties in areas of the city where building and therefore gentrification is occuring at rapid pace: Fishtown, Kensignton, Brewerytown, Point Breeze, Graduate Hospital, Queen's Village."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![10 Year Abatements 1](maps/Heat_Maps/Heat_Map_10yr_1.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![10 Year Abatements 1](maps/Heat_Maps/Heat_Map_10yr_2.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "ten_expt = opa_df[['exempt_building','homestead_exemption','lat','lng']].loc[opa_df['exempt_building'] > 0]\n",
    "ten_expt = ten_expt[['exempt_building','lat','lng']].loc[ten_expt['exempt_building'] != ten_expt['homestead_exemption']]\n",
    "ten_expt['lng'] = ten_expt['lng'].round(4)\n",
    "ten_expt['lat'] = ten_expt['lat'].round(4)\n",
    "grouped_ten = ten_expt.groupby(['lat','lng'])\n",
    "b = grouped_ten.count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "locations = b.index.values\n",
    "weights = b['exempt_building'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f23e34da90ac42aca5cda4e3770913df",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Figure(layout=FigureLayout(height='420px'))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Centered on City Hall\n",
    "fig = gmaps.figure(center = (39.9524, -75.1636), zoom_level=11, map_type = 'HYBRID')\n",
    "fig.add_layer(gmaps.heatmap_layer(locations, weights = weights, max_intensity = 100, point_radius = 5))\n",
    "fig"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Slider to show where new construction occuring in Philadelphia\n",
    "\n",
    "The 10 year tax abatement was enacted in 1997 to spur development in the city. The slider can be used to identify areas of the city and neighborhoods that have gone through development phases. The photos below are the new builds in 2017 and 2018, respectively. We can see in 2017-18 certain neighborhoods, Brewerytown, Fishtown, Kensignton are seeing new properties being built. While West Philly and Center City are not exepriencing the same building activity."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![Properties Build 2017](maps/Heat_Maps/Heat_Map_Slider_2017.png)\n",
    "![Properties Build 2018](maps/Heat_Maps/Heat_Map_Slider_2018.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\areuve200\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "age_df = opa_df[['year_built','lat','lng']]\n",
    "age_df['year_built'] = pd.to_numeric(age_df[\"year_built\"],errors='coerce')\n",
    "#remove nulls\n",
    "age_df = age_df.dropna()\n",
    "\n",
    "#remove values below 1980 and above 2018\n",
    "#1997 year of 10 year abatement enacted\n",
    "age_df = age_df.loc[(age_df['year_built'] >= 1980) & (age_df['year_built'] <= 2019)]\n",
    "age_df['lng'] = age_df['lng'].round(4)\n",
    "age_df['lat'] = age_df['lat'].round(4)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Code borrowed from https://jupyter-gmaps.readthedocs.io/en/latest/app_tutorial.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5c3412de4a2b4b7f901ce7c65da55915",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(HTML(value='<center><h3>Heatmap of properties built in Philadelphia, by year</h3></center>'), H…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from IPython.display import display\n",
    "import ipywidgets as widgets\n",
    "\n",
    "class PropertyExplorer(object):\n",
    "    \"\"\"\n",
    "    The user uses the slider to choose a year. This renders\n",
    "    a heatmap of properties built in that year.\n",
    "    \"\"\"\n",
    "\n",
    "    def __init__(self, df):\n",
    "        self._df = df\n",
    "        self._heatmap = None\n",
    "        self._slider = None\n",
    "        initial_year = min(self._df['year_built'])\n",
    "\n",
    "        title_widget = widgets.HTML(\n",
    "            '<center><h3>Heatmap of properties built in Philadelphia, by year</h3></center>'\n",
    "        )\n",
    "\n",
    "        map_figure = self._render_map(initial_year)\n",
    "        controls = self._render_controls(initial_year)\n",
    "        self._container = widgets.VBox([title_widget, controls, map_figure])\n",
    "\n",
    "    def render(self):\n",
    "        display(self._container)\n",
    "\n",
    "    def _on_year_change(self, change):\n",
    "        year = self._slider.value\n",
    "        self._heatmap.locations = self._locations_for_year(year)\n",
    "        self._total_box.value = self._total_properties_text_for_year(year)\n",
    "        return self._container\n",
    "\n",
    "    def _render_map(self, initial_year):\n",
    "        fig = gmaps.figure(map_type='HYBRID')\n",
    "        self._heatmap = gmaps.heatmap_layer(\n",
    "            self._locations_for_year(initial_year),\n",
    "            max_intensity=100,\n",
    "            point_radius=8\n",
    "        )\n",
    "        fig.add_layer(self._heatmap)\n",
    "        return fig\n",
    "\n",
    "    def _render_controls(self, initial_year):\n",
    "        self._slider = widgets.IntSlider(\n",
    "            value=initial_year,\n",
    "            min=min(self._df['year_built']),\n",
    "            max=max(self._df['year_built']),\n",
    "            description='Year',\n",
    "            continuous_update=False\n",
    "        )\n",
    "        self._total_box = widgets.Label(\n",
    "            value=self._total_properties_text_for_year(initial_year)\n",
    "        )\n",
    "        self._slider.observe(self._on_year_change, names='value')\n",
    "        controls = widgets.HBox(\n",
    "            [self._slider, self._total_box],\n",
    "            layout={'justify_content': 'space-between'}\n",
    "        )\n",
    "        return controls\n",
    "\n",
    "    def _locations_for_year(self, year):\n",
    "        return self._df[self._df['year_built'] == year][['lat', 'lng']]\n",
    "\n",
    "    def _total_properties_for_year(self, year):\n",
    "        return int(self._df[self._df['year_built'] == year]['year_built'].count())\n",
    "\n",
    "    def _total_properties_text_for_year(self, year):\n",
    "        return '{} Properties Built'.format(self._total_properties_for_year(year))\n",
    "\n",
    "\n",
    "PropertyExplorer(age_df[['year_built','lat','lng']]).render()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Machine Learning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Multiple Linear Regression\n",
    "\n",
    "Creating an multiple linear regression model to predict market value of residential real estate. Decided to limit the scope to residential for efficiency and for the possibility of better results. Further had to do some pre-processing to remove `NA` values as this method is unable to handle these values. Looked at residuals and the mean squared error (rmse) to assess the model performance with additional variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\areuve200\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\numpy\\core\\fromnumeric.py:2389: FutureWarning: Method .ptp is deprecated and will be removed in a future version. Use numpy.ptp instead.\n",
      "  return ptp(axis=axis, out=out, **kwargs)\n"
     ]
    }
   ],
   "source": [
    "opa_df_ols = opa_df[['number_of_bathrooms',\n",
    "                     'number_of_bedrooms',\n",
    "                     'total_livable_area',\n",
    "                     'market_value',\n",
    "                     'total_area',\n",
    "                     'depth',\n",
    "                     'number_of_rooms']].loc[opa_df['category_code'] == 1]\n",
    "opa_df_ols = opa_df_ols.dropna()\n",
    "X = opa_df_ols[['number_of_bathrooms','number_of_bedrooms','total_livable_area','total_area','depth','number_of_rooms']]\n",
    "X = sm.add_constant(X)\n",
    "y = opa_df_ols['market_value']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(323020, 7) (323020,)\n",
      "(138438, 7) (138438,)\n"
     ]
    }
   ],
   "source": [
    "# Create the training and test sets.\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.3, random_state = 6)\n",
    "# Check to see if the datasets split correctly.\n",
    "print(X_train.shape, y_train.shape)\n",
    "print(X_test.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>OLS Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>      <td>market_value</td>   <th>  R-squared:         </th>  <td>   0.085</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th>  <td>   0.085</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th>  <td>   4985.</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>             <td>Sat, 07 Sep 2019</td> <th>  Prob (F-statistic):</th>   <td>  0.00</td>   \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                 <td>11:09:22</td>     <th>  Log-Likelihood:    </th> <td>-4.6029e+06</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Observations:</th>      <td>323020</td>      <th>  AIC:               </th>  <td>9.206e+06</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Residuals:</th>          <td>323013</td>      <th>  BIC:               </th>  <td>9.206e+06</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Model:</th>              <td>     6</td>      <th>                     </th>      <td> </td>     \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>      <td>nonrobust</td>    <th>                     </th>      <td> </td>     \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "           <td></td>              <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>const</th>               <td>-2.628e+04</td> <td> 2125.833</td> <td>  -12.364</td> <td> 0.000</td> <td>-3.04e+04</td> <td>-2.21e+04</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>number_of_bathrooms</th> <td> 8.848e+04</td> <td> 1534.528</td> <td>   57.659</td> <td> 0.000</td> <td> 8.55e+04</td> <td> 9.15e+04</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>number_of_bedrooms</th>  <td>-2.806e+04</td> <td> 1043.022</td> <td>  -26.903</td> <td> 0.000</td> <td>-3.01e+04</td> <td> -2.6e+04</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_livable_area</th>  <td>  181.8569</td> <td>    1.359</td> <td>  133.801</td> <td> 0.000</td> <td>  179.193</td> <td>  184.521</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_area</th>          <td>   -2.9529</td> <td>    0.196</td> <td>  -15.049</td> <td> 0.000</td> <td>   -3.337</td> <td>   -2.568</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>depth</th>               <td>  -11.6944</td> <td>    2.213</td> <td>   -5.285</td> <td> 0.000</td> <td>  -16.031</td> <td>   -7.358</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>number_of_rooms</th>     <td>-1.038e+04</td> <td>  486.634</td> <td>  -21.324</td> <td> 0.000</td> <td>-1.13e+04</td> <td>-9423.347</td>\n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "  <th>Omnibus:</th>       <td>1776345.092</td> <th>  Durbin-Watson:     </th>      <td>   2.000</td>      \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Prob(Omnibus):</th>   <td> 0.000</td>    <th>  Jarque-Bera (JB):  </th> <td>176460146288792.250</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Skew:</th>            <td>304.573</td>   <th>  Prob(JB):          </th>      <td>    0.00</td>      \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Kurtosis:</th>      <td>114503.685</td>  <th>  Cond. No.          </th>      <td>1.41e+04</td>      \n",
       "</tr>\n",
       "</table><br/><br/>Warnings:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.<br/>[2] The condition number is large, 1.41e+04. This might indicate that there are<br/>strong multicollinearity or other numerical problems."
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                            OLS Regression Results                            \n",
       "==============================================================================\n",
       "Dep. Variable:           market_value   R-squared:                       0.085\n",
       "Model:                            OLS   Adj. R-squared:                  0.085\n",
       "Method:                 Least Squares   F-statistic:                     4985.\n",
       "Date:                Sat, 07 Sep 2019   Prob (F-statistic):               0.00\n",
       "Time:                        11:09:22   Log-Likelihood:            -4.6029e+06\n",
       "No. Observations:              323020   AIC:                         9.206e+06\n",
       "Df Residuals:                  323013   BIC:                         9.206e+06\n",
       "Df Model:                           6                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "=======================================================================================\n",
       "                          coef    std err          t      P>|t|      [0.025      0.975]\n",
       "---------------------------------------------------------------------------------------\n",
       "const               -2.628e+04   2125.833    -12.364      0.000   -3.04e+04   -2.21e+04\n",
       "number_of_bathrooms  8.848e+04   1534.528     57.659      0.000    8.55e+04    9.15e+04\n",
       "number_of_bedrooms  -2.806e+04   1043.022    -26.903      0.000   -3.01e+04    -2.6e+04\n",
       "total_livable_area    181.8569      1.359    133.801      0.000     179.193     184.521\n",
       "total_area             -2.9529      0.196    -15.049      0.000      -3.337      -2.568\n",
       "depth                 -11.6944      2.213     -5.285      0.000     -16.031      -7.358\n",
       "number_of_rooms     -1.038e+04    486.634    -21.324      0.000   -1.13e+04   -9423.347\n",
       "===============================================================================\n",
       "Omnibus:                  1776345.092   Durbin-Watson:                    2.000\n",
       "Prob(Omnibus):                  0.000   Jarque-Bera (JB):   176460146288792.250\n",
       "Skew:                         304.573   Prob(JB):                          0.00\n",
       "Kurtosis:                  114503.685   Cond. No.                      1.41e+04\n",
       "===============================================================================\n",
       "\n",
       "Warnings:\n",
       "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
       "[2] The condition number is large, 1.41e+04. This might indicate that there are\n",
       "strong multicollinearity or other numerical problems.\n",
       "\"\"\""
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Build a linear model on training data\n",
    "model_m = sm.OLS(y_train, X_train).fit()\n",
    "model_m.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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>actual</th>\n",
       "      <th>predicted</th>\n",
       "      <th>residuals</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>395734</th>\n",
       "      <td>203800.0</td>\n",
       "      <td>225826.477081</td>\n",
       "      <td>-22026.477081</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>454233</th>\n",
       "      <td>77900.0</td>\n",
       "      <td>119090.296257</td>\n",
       "      <td>-41190.296257</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>536694</th>\n",
       "      <td>282600.0</td>\n",
       "      <td>178476.398417</td>\n",
       "      <td>104123.601583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>177270</th>\n",
       "      <td>69300.0</td>\n",
       "      <td>94519.734843</td>\n",
       "      <td>-25219.734843</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>321412</th>\n",
       "      <td>88700.0</td>\n",
       "      <td>99194.449776</td>\n",
       "      <td>-10494.449776</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          actual      predicted      residuals\n",
       "395734  203800.0  225826.477081  -22026.477081\n",
       "454233   77900.0  119090.296257  -41190.296257\n",
       "536694  282600.0  178476.398417  104123.601583\n",
       "177270   69300.0   94519.734843  -25219.734843\n",
       "321412   88700.0   99194.449776  -10494.449776"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Predict values of `market_value` using the test data.\n",
    "actual = y_test\n",
    "prediction = model_m.predict(X_test)\n",
    "residuals = actual - prediction\n",
    "\n",
    "property_result = pd.concat([actual.rename('actual'),prediction.rename('predicted'),residuals.rename('residuals')], axis = 1)\n",
    "property_result.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "158868.1017592656\n"
     ]
    }
   ],
   "source": [
    "print(rmse(prediction,actual))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This portion of code below would remove outliers, however, after a few hours of running it does not return results. I got it to run on a data set of significanlty less recods, ~19k. This is a system limitation and ideally we would remove outliers, re-run the model and then predict the `market_value` based on the best OLS model.\n",
    "\n",
    "```python\n",
    "test_m = model_m.outlier_test()\n",
    "test_final_m = test_m[test_m['bonf(p)'] < 0.05]\n",
    "\n",
    "# Drop outliers from both X and y train sets.\n",
    "X_train_no_outliers = X_train.drop(test_final_m.index)\n",
    "y_train_no_outliers = y_train.drop(test_final_m.index)\n",
    "\n",
    "model_m_no_outliers =sm.OLS(y_train_no_outliers, X_train_no_outliers).fit()\n",
    "prediction = model_m_no_outliers.predict(X_test)\n",
    "\n",
    "```"
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}