Predict_accessibility_using_ML.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predict park accessibility in NYC using Machine Learning\n",
    "\n",
    "In this notebook, we work through how we can ingest the parquet file produced by our ETL pipeline into an ML algorithm for training a model to predict accessibility. We will go through the entire data science lifecyle -- \n",
    "\n",
    "data ingest --> data exploration --> data cleaning --> staging the training data --> machine learning.\n",
    "\n",
    "Accessiblity here is defined as whether parks are categorized as Level 4 or not.  \n",
    "\n",
    "For Machine learning, we will use two approaches -- \n",
    "\n",
    "1/ Train an XGBoost model to predict whether accessible parks are available based on individual tax returns and park data \n",
    "\n",
    "2/ Train a model using SageMaker Autopilot and compare the results. SageMaker Autopilot is a fully managed ML service that can train various machine learning models with different hyperparameters, and allow you to pick the best one. Here we will use the SageMaker Python SDK to make API calls to SageMaker Autopilot to train and deploy an ML model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Import Libraries and Data\n",
    "\n",
    "Note that to import the data in parquet format into a pandas dataframe, we will need to install pyarrow library."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.metrics import accuracy_score, auc, classification_report\n",
    "from sklearn.model_selection import train_test_split\n",
    "import io\n",
    "import sagemaker.amazon.common as smac\n",
    "import boto3\n",
    "import re\n",
    "from sagemaker import get_execution_role\n",
    "import sagemaker\n",
    "from sagemaker.amazon.amazon_estimator import get_image_uri\n",
    "from sagemaker.predictor import csv_serializer, json_deserializer\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already up-to-date: pip in /home/ec2-user/anaconda3/envs/python3/lib/python3.6/site-packages (19.3.1)\r\n"
     ]
    }
   ],
   "source": [
    "# if needed, updated boto3\n",
    "!pip install --upgrade pip\n",
    "if boto3.__version__ != '1.10.33':\n",
    "    !pip install boto3=='1.10.33'\n",
    "else:\n",
    "    pass\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: pyarrow in /home/ec2-user/anaconda3/envs/python3/lib/python3.6/site-packages (0.15.1)\r\n",
      "Requirement already satisfied: numpy>=1.14 in /home/ec2-user/anaconda3/envs/python3/lib/python3.6/site-packages (from pyarrow) (1.14.3)\r\n",
      "Requirement already satisfied: six>=1.0.0 in /home/ec2-user/anaconda3/envs/python3/lib/python3.6/site-packages (from pyarrow) (1.11.0)\r\n"
     ]
    }
   ],
   "source": [
    "!pip install pyarrow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "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>Accessible</th>\n",
       "      <th>Adaptive_Swing</th>\n",
       "      <th>Level</th>\n",
       "      <th>Playground_ID</th>\n",
       "      <th>School_ID</th>\n",
       "      <th>Status</th>\n",
       "      <th>lat</th>\n",
       "      <th>lon</th>\n",
       "      <th>num_of_returns</th>\n",
       "      <th>zipcode</th>\n",
       "      <th>...</th>\n",
       "      <th>num_of_head_of_household_returns</th>\n",
       "      <th>num_of_volunteer_prepared_returns_Num_of_volunteer_income_tax_assistance_prepared_returns</th>\n",
       "      <th>num_of_exemptions</th>\n",
       "      <th>Name</th>\n",
       "      <th>num_of_volunteer_prepared_returns_Num_of_tax_counseling_for_the_elderly_prepared_returns</th>\n",
       "      <th>Prop_ID</th>\n",
       "      <th>Location</th>\n",
       "      <th>num_of_volunteer_prepared_returns_Total</th>\n",
       "      <th>num_of_joint_returns</th>\n",
       "      <th>num_of_dependents</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Y</td>\n",
       "      <td>N</td>\n",
       "      <td>4</td>\n",
       "      <td>B080-01</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>40.5753</td>\n",
       "      <td>-73.9722</td>\n",
       "      <td>10120.0</td>\n",
       "      <td>11224.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2310.0</td>\n",
       "      <td>410.0</td>\n",
       "      <td>16270.0</td>\n",
       "      <td>Asser Levy Park</td>\n",
       "      <td>0.0</td>\n",
       "      <td>B080</td>\n",
       "      <td>Boardwalk, Surf Ave., Sea Breeze Ave., Ocean P...</td>\n",
       "      <td>410.0</td>\n",
       "      <td>1380.0</td>\n",
       "      <td>5480.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Y</td>\n",
       "      <td>N</td>\n",
       "      <td>4</td>\n",
       "      <td>B080-01</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>40.5753</td>\n",
       "      <td>-73.9722</td>\n",
       "      <td>4890.0</td>\n",
       "      <td>11224.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1510.0</td>\n",
       "      <td>130.0</td>\n",
       "      <td>9100.0</td>\n",
       "      <td>Asser Levy Park</td>\n",
       "      <td>0.0</td>\n",
       "      <td>B080</td>\n",
       "      <td>Boardwalk, Surf Ave., Sea Breeze Ave., Ocean P...</td>\n",
       "      <td>130.0</td>\n",
       "      <td>960.0</td>\n",
       "      <td>3260.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Y</td>\n",
       "      <td>N</td>\n",
       "      <td>4</td>\n",
       "      <td>B080-01</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>40.5753</td>\n",
       "      <td>-73.9722</td>\n",
       "      <td>2350.0</td>\n",
       "      <td>11224.0</td>\n",
       "      <td>...</td>\n",
       "      <td>580.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4320.0</td>\n",
       "      <td>Asser Levy Park</td>\n",
       "      <td>0.0</td>\n",
       "      <td>B080</td>\n",
       "      <td>Boardwalk, Surf Ave., Sea Breeze Ave., Ocean P...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>630.0</td>\n",
       "      <td>1350.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Y</td>\n",
       "      <td>N</td>\n",
       "      <td>4</td>\n",
       "      <td>B080-01</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>40.5753</td>\n",
       "      <td>-73.9722</td>\n",
       "      <td>1280.0</td>\n",
       "      <td>11224.0</td>\n",
       "      <td>...</td>\n",
       "      <td>260.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2490.0</td>\n",
       "      <td>Asser Levy Park</td>\n",
       "      <td>0.0</td>\n",
       "      <td>B080</td>\n",
       "      <td>Boardwalk, Surf Ave., Sea Breeze Ave., Ocean P...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>520.0</td>\n",
       "      <td>700.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Y</td>\n",
       "      <td>N</td>\n",
       "      <td>4</td>\n",
       "      <td>B080-01</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>40.5753</td>\n",
       "      <td>-73.9722</td>\n",
       "      <td>1350.0</td>\n",
       "      <td>11224.0</td>\n",
       "      <td>...</td>\n",
       "      <td>180.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3100.0</td>\n",
       "      <td>Asser Levy Park</td>\n",
       "      <td>0.0</td>\n",
       "      <td>B080</td>\n",
       "      <td>Boardwalk, Surf Ave., Sea Breeze Ave., Ocean P...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>810.0</td>\n",
       "      <td>950.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  Accessible Adaptive_Swing Level Playground_ID School_ID Status      lat  \\\n",
       "0          Y              N     4       B080-01      None   None  40.5753   \n",
       "1          Y              N     4       B080-01      None   None  40.5753   \n",
       "2          Y              N     4       B080-01      None   None  40.5753   \n",
       "3          Y              N     4       B080-01      None   None  40.5753   \n",
       "4          Y              N     4       B080-01      None   None  40.5753   \n",
       "\n",
       "        lon  num_of_returns  zipcode  ...  num_of_head_of_household_returns  \\\n",
       "0  -73.9722         10120.0  11224.0  ...                            2310.0   \n",
       "1  -73.9722          4890.0  11224.0  ...                            1510.0   \n",
       "2  -73.9722          2350.0  11224.0  ...                             580.0   \n",
       "3  -73.9722          1280.0  11224.0  ...                             260.0   \n",
       "4  -73.9722          1350.0  11224.0  ...                             180.0   \n",
       "\n",
       "   num_of_volunteer_prepared_returns_Num_of_volunteer_income_tax_assistance_prepared_returns  \\\n",
       "0                                              410.0                                           \n",
       "1                                              130.0                                           \n",
       "2                                                0.0                                           \n",
       "3                                                0.0                                           \n",
       "4                                                0.0                                           \n",
       "\n",
       "   num_of_exemptions             Name  \\\n",
       "0            16270.0  Asser Levy Park   \n",
       "1             9100.0  Asser Levy Park   \n",
       "2             4320.0  Asser Levy Park   \n",
       "3             2490.0  Asser Levy Park   \n",
       "4             3100.0  Asser Levy Park   \n",
       "\n",
       "   num_of_volunteer_prepared_returns_Num_of_tax_counseling_for_the_elderly_prepared_returns  \\\n",
       "0                                                0.0                                          \n",
       "1                                                0.0                                          \n",
       "2                                                0.0                                          \n",
       "3                                                0.0                                          \n",
       "4                                                0.0                                          \n",
       "\n",
       "   Prop_ID                                           Location  \\\n",
       "0     B080  Boardwalk, Surf Ave., Sea Breeze Ave., Ocean P...   \n",
       "1     B080  Boardwalk, Surf Ave., Sea Breeze Ave., Ocean P...   \n",
       "2     B080  Boardwalk, Surf Ave., Sea Breeze Ave., Ocean P...   \n",
       "3     B080  Boardwalk, Surf Ave., Sea Breeze Ave., Ocean P...   \n",
       "4     B080  Boardwalk, Surf Ave., Sea Breeze Ave., Ocean P...   \n",
       "\n",
       "   num_of_volunteer_prepared_returns_Total num_of_joint_returns  \\\n",
       "0                                    410.0               1380.0   \n",
       "1                                    130.0                960.0   \n",
       "2                                      0.0                630.0   \n",
       "3                                      0.0                520.0   \n",
       "4                                      0.0                810.0   \n",
       "\n",
       "  num_of_dependents  \n",
       "0            5480.0  \n",
       "1            3260.0  \n",
       "2            1350.0  \n",
       "3             700.0  \n",
       "4             950.0  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "FILENAME = 'part-00000-588c7835-a113-419a-8cea-580b11e397e8-c000.snappy.parquet'\n",
    "df = pd.read_parquet(FILENAME)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Q133        12\n",
       "B111        12\n",
       "B066        12\n",
       "Q104-01     12\n",
       "M041         6\n",
       "Q162E        6\n",
       "B332         6\n",
       "X168         6\n",
       "Q300-05      6\n",
       "Q205A        6\n",
       "X171         6\n",
       "M113         6\n",
       "B306         6\n",
       "B121         6\n",
       "X034-01      6\n",
       "M058-06      6\n",
       "Q123         6\n",
       "M243         6\n",
       "B164         6\n",
       "Q426         6\n",
       "X300         6\n",
       "M200         6\n",
       "M184         6\n",
       "B135         6\n",
       "M308         6\n",
       "X194         6\n",
       "X258         6\n",
       "X116         6\n",
       "R067         6\n",
       "X038-01      6\n",
       "            ..\n",
       "Q163-21      1\n",
       "Q099-15      1\n",
       "X010-02      1\n",
       "X010-05      1\n",
       "Q020-02      1\n",
       "M071-18      1\n",
       "Q027         1\n",
       "B057-04      1\n",
       "Q015-05      1\n",
       "Q001-01      1\n",
       "X010-03      1\n",
       "M399-01      1\n",
       "Q458         1\n",
       "Q008-01      1\n",
       "Q300A-01     1\n",
       "M283-01      1\n",
       "Q300B        1\n",
       "M010-176     1\n",
       "X039-01      1\n",
       "B058-01      1\n",
       "X010-08      1\n",
       "M010-202     1\n",
       "M037-04      1\n",
       "X030-99      1\n",
       "Q008-02      1\n",
       "B028-01      1\n",
       "Q021-01      1\n",
       "Q162G13      1\n",
       "M140-01      1\n",
       "M010-145     1\n",
       "Name: Playground_ID, Length: 1016, dtype: int64"
      ]
     },
     "execution_count": 133,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.Playground_ID.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of Dataset = (5423, 23)\n"
     ]
    }
   ],
   "source": [
    "print(\"Shape of Dataset = {}\".format(df.shape))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Accessible',\n",
       " 'Adaptive_Swing',\n",
       " 'Level',\n",
       " 'Playground_ID',\n",
       " 'School_ID',\n",
       " 'Status',\n",
       " 'lat',\n",
       " 'lon',\n",
       " 'num_of_returns',\n",
       " 'zipcode',\n",
       " 'num_of_single_returns',\n",
       " 'size_of_adjusted_gross_income',\n",
       " 'num_with_paid_preparers_signature',\n",
       " 'num_of_head_of_household_returns',\n",
       " 'num_of_volunteer_prepared_returns_Num_of_volunteer_income_tax_assistance_prepared_returns',\n",
       " 'num_of_exemptions',\n",
       " 'Name',\n",
       " 'num_of_volunteer_prepared_returns_Num_of_tax_counseling_for_the_elderly_prepared_returns',\n",
       " 'Prop_ID',\n",
       " 'Location',\n",
       " 'num_of_volunteer_prepared_returns_Total',\n",
       " 'num_of_joint_returns',\n",
       " 'num_of_dependents']"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(df.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Cleaning prior to Exploration"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are several data cleaning steps that need to be performed:\n",
    "\n",
    "1) There isn't enough data to identify each playground Id separately but we can extract the Borough information as a categorical variable <br/>\n",
    "2) We need to convert the AGI to ordinal values <br/>\n",
    "3) We need to convert Accessible and Adaptible Swing to numeric values <br/>\n",
    "4) We will drop the name and location columns <br/>\n",
    "5) We will drop any columns on volunteer prepared taxes <br/>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Borough'] = [str(x)[0] for x in df.Playground_ID.values]\n",
    "convert_cols = {'Accessible': {'Y': 1, 'N': 0},\n",
    "                'Adaptive_Swing': {'Y': 1, 'N': 0}}#,\n",
    "              # 'adjusted_gross_income': {'':0, '$1 under $25,000':1, '$25,000 under $50,000': 2, '$50,000 under $75,000':3, \n",
    "              #                          '$75,000 under $100,000':4, '$100,000 under $200,000':5, '$200,000 or more':6}}\n",
    "df.replace(convert_cols, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ec2-user/anaconda3/envs/python3/lib/python3.6/site-packages/pandas/core/ops.py:1649: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
      "  result = method(y)\n"
     ]
    }
   ],
   "source": [
    "COLSCONVERT = ['Adaptive_Swing', 'num_of_exemptions',\n",
    "       'Accessible', 'num_of_dependents', 'num_of_returns',\n",
    "       'num_of_joint_returns', 'num_of_head_of_household_returns',\n",
    "       'num_of_single_returns']\n",
    "indexlist = []\n",
    "for col in COLSCONVERT:\n",
    "    indexlist = indexlist + list(df.loc[df[col]=='**'].index)\n",
    "df = df.drop(index= indexlist)\n",
    "df[COLSCONVERT] = df[COLSCONVERT].apply(pd.to_numeric)\n",
    "df.replace({'Level': {None:0, '1':0, '2':0, '3':0, '4':1}}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "COLS_TO_DROP = ['lat', 'lon', 'Playground_ID', 'Name', 'Location', 'Prop_ID','num_of_volunteer_prepared_returns_Total',\n",
    "                  'num_of_volunteer_prepared_returns_Num_of_volunteer_income_tax_assistance_prepared_returns',\n",
    "                 'num_with_paid_preparers_signature','num_of_volunteer_prepared_returns_Num_of_tax_counseling_for_the_elderly_prepared_returns', \n",
    "                'zipcode','School_ID','Status']\n",
    "df = df.drop(columns = COLS_TO_DROP)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "# let's save this dataframe at this stage as a separate dataframe for AutoPilot\n",
    "autodf = df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Mark everything with Level 4 as 1, and else 0. Why have we done this?\n",
    "\n",
    "Look at the graph below. It shows that most playgrounds in NYC are classified as Level 4 playgrounds.  Level 4 means Accessible Playgrounds with Transfer Platforms and Ground Level Play Features; and thus represents the most sophisticated playgrounds.\n",
    "\n",
    "We want to predict whether a given Playground ID in a given school district, given demographic information about the income levels of households in that zipcode is a Level 4 playground or not. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8931a9bdd8>"
      ]
     },
     "execution_count": 139,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(df.Level)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Baseline Model accuracy = 0.6511156186612576\n"
     ]
    }
   ],
   "source": [
    "print(\"Baseline Model accuracy = {}\".format(len(df[df['Level']==1])/len(df)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4590, 11)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.dropna(inplace=True)\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Exploration\n",
    "\n",
    "Having converted the raw data into numerical format, we can explore the dataset and extract some meaningful information and look for any correlations in the data, which may be indicative of whether income is correlated to playground type"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Total returns versus AGI shows an overall trend that we need to normalize out. \n",
    "sns.barplot(x='size_of_adjusted_gross_income', y = 'num_of_returns', data = df)\n",
    "plt.title(\"Total Returns versus AGI\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Number of dependents versus number of exceptions\n",
    "sns.regplot(x = 'num_of_dependents', y = 'num_of_exemptions', data = df)\n",
    "plt.title(\"Scatter plot of dependents versus exceptions\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Introduce new columns which are the normalized number of joint, head of household and single returns\n",
    "COLS_TO_NORMALIZE = ['num_of_joint_returns', 'num_of_head_of_household_returns',\n",
    "       'num_of_single_returns']\n",
    "for col in COLS_TO_NORMALIZE:\n",
    "    df['normalized_'+ col] = df[col]/df['num_of_returns']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(x='size_of_adjusted_gross_income', y = 'normalized_num_of_head_of_household_returns', data = df)\n",
    "plt.title(\"AGI versus normalized HOH returns\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(x='Accessible', y = 'num_of_dependents', hue = 'Borough', data = df)\n",
    "plt.title(\"Playground Accessibility by number of dependents and Borough\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.boxplot(x = 'Accessible', y = 'normalized_num_of_joint_returns', hue = 'Borough', data = df)\n",
    "plt.title(\"Playground accessibility by normalized number of joint returns\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Not much correlation between adjusted gross income level and number of available playgrounds. Certain zip codes simply\n",
    "# dont have any filers in the high income categories. \n",
    "sns.countplot(x = 'Level' , hue = 'size_of_adjusted_gross_income', data = df)\n",
    "plt.title(\"Distribution of Accessibility Levels by Income Accross all Boroughs\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# How does this play out at a Borough specific level? Let's look at Manhattan and Bronx\n",
    "bordf = df[df['Borough']=='X']\n",
    "sns.countplot(x = 'Level', hue = 'size_of_adjusted_gross_income', data = bordf)\n",
    "plt.title(\"Distribution of Accessibility Levels by Income in the Bronx\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(x = 'size_of_adjusted_gross_income', hue = 'Level', data = bordf)\n",
    "plt.title(\"Distribution of Accessibility Levels by Income in the Manhattan\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAEWCAYAAACufwpNAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3XmcVnP/x/HXp6nUpEWLaLmVG91EhRYKlbRKC5XcpCi7Qiqkn5tu4kaW25IbIXuKypJkKd3WFqTUTbY0MsmkRUoz0+f3xzkzZqZZrqlzzVzV+/l4zGOu65zv+Z7PWa7zOed7zvW9zN0RERGJQpnSDkBERPYcSioiIhIZJRUREYmMkoqIiERGSUVERCKjpCIiIpEp8aRiZg+Z2f9FVNdfzOw3M0sK3881syFR1B3W97qZDYyqvmLM92Yz+8XMUkt63lExsy/MrF34+kYze7q4ZfNu391NUctdyHTfm9kp8Yhpd2JmZ5vZ7Ajri/T4EG87u/+UtkiTSvhh2GJmm8xsvZl9YGYXm1n2fNz9Ynf/Z4x1FfrBcvcf3H1fd8+MIPYdNqC7d3X3SbtadzHj+AtwNXCEux9QSLmGZrbdzCaUXHSxc/fG7j53V8rm3b67elAwMzezQ3Z2+t2NmbUzs5TSjmNnufsz7t5pZ6aN9wE5PD5tM7OaeYZ/Gu5nDeI171iY2SAzey/PsCfM7OZ4zzseVyqnuXtl4CDgNuAaYGLUMzGzslHXmSD+AqS5+89FlDsX+BU408z2iX9YIpLHd8BZWW/M7CggufTCSRDuHtkf8D1wSp5hLYHtwJHh+yeAm8PXNYFXgfXAOuC/BInuqXCaLcBvwCigAeDAYOAHYF6OYWXD+uYCtwLzgY3ADKB6OK4dkJJfvEAXYBuQHs5vcY76hoSvywBjgJXAz8CTQNVwXFYcA8PYfgGuL2Q9VQ2nXxvWNyas/5RwmbeHcTxRwPQGfANcAqwB+uQZ3xh4M1yna4DR4fAkYHQ47SZgEVA/HPe3HNN8CfTLUV83YFk4zY/AiMK2X959AbgRmApMDuv4BGia334Tln06z3otC9wCZAJbw3VzP/AAMD7Psr8MXFXAenPgkALGnQ8sJ0jUbwAHhcMnAHfmKTsDGB6+rgO8GG7L74BhOcrlXJYKwNNAWri+FgC1C/kcXReu81+Bx4EK4bilBCduWWXLEexvR+dTTzty7PME+/M/gffD7TAbqJlj/AnAB2F8q4BBhe2v4bhBYX13h9N9C7QOh68i+KwMzDGPfYA7CT4na4CHgIoFrIdBwHt5tt/FwIpwXg8Als90hX2eC1v+43Is/2KgXRHHujHAghzD7gSuD+NsEA47FfiU4Hi0CrgxR/kGFHLcCPefF8J1vwn4AmieY/y1/PlZXgb0DocfTvA5yQyXfz1wYbg+toXDXimsjpzrP1yuXwn2765F5oGiChTnj3ySSjj8B+CS8PUT/JlUbg13qnLh34lZO0neunJsgCeBSkBF8k8qPwJHhmVe5M8PdTsKSCp5DwB5PoRZSeV84GvgYGBf4CXgqTyxPRLG1RT4Azi8gPX0JMGBqXI47VfA4ILizGf6E8P69wPuy9pBwnGVgZ8ImtAqhO9bheNGAkuARgSJqSlQI1xXq4DzCA7gRxPs4EeE0/0EnBi+3g84pjjbL1y36UCfsNwIgh20XGHboYDtOyTHsrYEVvPnAa4m8DsFH6zzTSpAz3DbHh4u/xjgg3DcSeG6sRzLv4UgmZQhSMw3AOXDfeNboHM+y3IR8ArBmWwScCxQpZDP0VKgPlCd4CCY9ZkZBUzOE/uSAuppx45J5RvgMIL9dC5wWzjuIIIDy1nhNqoBNIthfx0EZBDsO0nAzQSf9wcIEkinsN59w/J3EyT+6mF9rwC3FhD/IHZMKq8C1Qiu6NcCXQqYNnvdx7j8dQkSfrdwu3YM39cq7FhHcAJ2eLjsKeF6zJlU2gFHhXU2IUikvWI5boTLsDWMKYng8/ZRjhj68ud+eCawGTgwv3WX99hbjDrSgQvC+V9C8HnbIZHn/CupG/WrCXaivNKBAwnOCtPd/b8eLk0hbnT3ze6+pYDxT7n7UnffDPwf0C+iG71nA3e5+7fu/hvBmWT/PM1wN7n7FndfTHCm0zRvJWEs/YHr3H2Tu38PjAcGFCOWgcDr7v4r8CzQxcz2D8d1B1Ldfby7bw3n8XE4bggwxt2/9MBid08Lp/ne3R939wx3/5QgIfcNp0sHjjCzKu7+q7t/kmN4rNtvkbtPdfd04C6ChHdcMZZ5B+4+H9gAdAgH9QfmuvuaYlZ1McGBbbm7ZwDjgGZmdhDB1ZcTJEwIEuOH7r4aaEFw0Bnr7tvc/VuCA0T/fOaRTnCgPsTdM919kbtvLCSm+919lbuvI7hKy2pmeRroZmZVwvcDCK7sY/W4u38Vfn5eAJqFw/8OvOXuz4XbMs3dP4txf/0u3HcyCa5G6wNj3f0Pd59NcHZ8iJkZwRnzVe6+zt03Eazr/NZXQW5z9/Xu/gMwJ0f8u7r85wAz3X2mu2939zeBhQQH9MI8RdAU3ZHgSvfHnCPdfa67Lwnr/Bx4Dmibp47CjhvvhTFlhvPKHufuU9x9dVj3ZIIruJYxr4nY6ljp7o+E859E8HmvXVidJZVU6hI0j+R1B8EZ4mwz+9bMro2hrlXFGL+S4KyrZgFli6NOWF/OusuSewXnfFrrd4IrmrxqhjHlratuLEGYWUWCg/0zAO7+IcGZ4d/DIvUJzsbyU9C4g4BW4cMV681sPUESzXpQ4AyCD9dKM3vXzI4Phxdn+2VvF3ffTnBWV6fQhY3NJIIDAuH/4hxgsxwE3Jtj2dcRXMnVDZPk8/x5UP874boPp6uTZ72NJv8P3VMEzWrPm9lqM7vdzMoVElPe/bgOQJjM3gfOMLNqQNcc8cSioH20oH0jlv01ZxLfEsaZd9i+QC2CK7VFOdbXrHD4rsa/q9MfBPTNsy1PIDiIFuYpgn1iEMEVXS5m1srM5pjZWjPbQHACk/d4VNgy5R1XIetE1szONbPPcsR7ZD51FyqGOrLn7+6/hy8LXedxTypm1oJgB3wv77jwzOdqdz8Y6AEMN7Oss86CzniLupKpn+P1XwjOEH8huKzLvokWnoHl3JmLqnc1wY6Xs+4Mcn+gYvFLGFPeun7Mv/gOegNVgAfNLDV87LguwdULBAejgwuYdhXw1wKGv+vu1XL87evulwC4+wJ37wnsD0wnOMMravvllb1dwqcB6xGs0+LIbxs9DfQ0s6YEzRDTi1knBMt/UZ7lr+juH4TjnwP6hFcurQiu4rKm+y7PdJXdfYez2/Ds/yZ3P4LgnkN3gjPcguTdj3Ouq6xE2pfgqinWfacwBe0bu7q/5q1rC9A4x/qq6u7FTQyxKOrznNcqglaOnNuykrvfVuhM3FcSNOV2I2gSz+tZgua++u5elaC52IoZ2w7CffER4HKghrtXI2gyzao7v+XPNSyGOnZK3JKKmVUxs+4EZ3lPu/uSfMp0N7Osy+INBDeWtoej11DwwbEw55jZEWaWDIwFpoaXbl8RZPlTwzPEMQRtvlnWAA1yPv6cx3PAVeGjvPsSXLZPDptLYhbG8gJwi5lVDjfscIKDYywGAo8RtNM2C//aAE3Dp09eBQ40syvNbJ9wHq3CaR8F/mlmh1qgiZnVCKc5zMwGmFm58K+FmR1uZuUt+L5A1bDpaiPhNipi++V1rJmdHp5lXUnQdvxR7GsOyGefcPcUgpveTwEvesHNolnKm1mFHH9JBB/068yscbhcVc0sq+mPsDnwF4L194a7rw9HzQc2mdk1ZlbRzJLM7MjwRCoXM2tvZkeF89tIcKAuaF0BXGZm9cysOsHN38k5xk0HjgGuIJ+z4530DHCKmfUzs7JmVsPMmkWwv2YLr1AfAe7Oaq41s7pm1jmiZcipqM9zXk8Dp5lZ53A7VrDgkex6MUw7GDjZgyb3vCoD69x9q5m15M8WhV1ViSBJrAUws/MIrjKyrAHqmVn5PMNyfn6KqmOnxCOpvGJmmwgy//UE7efnFVD2UOAtgqcRPgQedPc54bhbgTHhZdmIYsz/KYIbUqkE7fbDANx9A3ApwYHhR4Irl5zP8E8J/6eZ2Sfs6LGw7nkEZyZbgaHFiCunoeH8vyW4gns2rL9QZlaX4P7BPe6emuNvEUEzwsCwnbojcBrBOlgBtA+ruIvgADGb4MA2keDJm00EN1T7E5wRpwL/4s+kOwD43sw2Ely+nx0OL2z75TWD4Ebgr2F9p4dJqjjuJbhi+NXM/p1j+CSCJBtL09cXBGfLWX/nufs0guV9PlzGpQTNSjk9S3Bj9tmsAeEBtztBYv+OPxNP1XzmewDBE3AbCdre3y0i3mcJttO3BM1S2d8vCBPni0BD8j87LrbwHkU3ggc81gGf8Wf7/U7trwW4hqDJ9KNwXb9F8OBI1Ir6POfi7qsIHnoYTXCQXUXwYEuRx0h3/8bdFxYw+lJgbHhMvIHwKn9XufsygntbHxIki6MImkWzvEOwr6ea2S/hsIkE90bXm9n0GOrYKVlPtIjstszsJIIzzYN8L9mhzewG4DB3P6fIwiIlaE/9AqHsJcKmzCuAR/eihFKdoMmlOE8MipQIdSgpuy0zO5zgi10HAveUcjglwswuIGiaed3d55V2PCJ5qflLREQioysVERGJzG5xT6VmzZreoEGD0g5DRGS3smjRol/cvThfLt1lu0VSadCgAQsXFvTEnoiI5MfMVhZdKlpq/hIRkcgoqYiISGSUVEREJDK7xT0VkUSUnp5OSkoKW7duLe1QZC9XoUIF6tWrR7lyhXV8XTKUVER2UkpKCpUrV6ZBgwYEfWqKlDx3Jy0tjZSUFBo2bFja4aj5S2Rnbd26lRo1aiihSKkyM2rUqJEwV8xKKiK7QAlFEkEi7YdKKiIiEhklFRERiYxu1IvE0ZAhQxg+fDhHHHFEgWWWrfqlwHGx+PbrFYy4/ALMjLsnPMZfGgQ3a/+aVLxfun5y8nQ++fwL7rnleh55cjIVK1bgnL498y37XVJSvsM3btjIqy+9yt/PK94PHN5/x/0kV0rm/846v1jTxWpLpQN2afp/jR/HcS1b0/bEdjuMy0xP26W6C3Lgwfn9unPiU1IRiaNHH3007vN4+42ZdOp2GhcPuzqyOi8498ydmm7Thk0898RzxU4qUcjIyKBs2fgc0q65enRc6t0TqflLJCKbN2/m1FNPpWnTphx55JFMnjyZdu3asXDhQl5++WWaNWtGs2bNaNSoUfajn4sWLWJg3x707daBC87py9o1qQXWv/yLJZzVswu9O7Vl2AUD2bB+PfPeeZOnHvsPk596gkFn9ipw2r7nD+P4Lv04un1PHn16SvbwSZOnceQJp3LCqf35cOGn2cP/Of4B7n7ocQA69hnEosVLAfhl3a8c1qoTACv+t4J+XfrRu0Nverbvyffffs/4W8azauUqenfozR033QHAxAcm0rdzX3q278l9t9+XPY+H7nmILq27cHaPs/num+8KXbeLlizh+J49adO7N2PuuINWp50GwDPTpnHmpZfSfdAgTjvvPNw9e/xxPXrw4syZAKxZk0rPPt04ucuJnHTK8Xz08QdkZmYybPilnHTK8bTt2JqHHn2wwPkPG34pr7w2A4DmrZtw+/hbOaVbW9p2bM2Kb74Bgu1/5ahrOLlrNzp0O5XXZs0CYNrLr3By126079KVm/91e3adhxzVhLG33ka7Ll3oN+BcPl28mDP+/neOa9eeN956C4DMzExGjhxJixYtaNKkCf/5z38KXU+JQFcqIhGZNWsWderU4bXXXgNgw4YNTJgwAYAePXrQo0cPAPr160fbtm1JT09n6NCh3P3QY1SvUZPXX57GvXeM4+Y7/51v/aOvupzRY8fR4rg23Df+Nh685w6uu/EWzjx7EMmVKnHeRZcVGNt/xv+T6vtVZcuWrbQ5tT+9u3VkW3o6/7zzAT6c9QJVK1emU9/zaHbk4TEv7+QnJzPgggGcdsZpbNu2je2Z27n6+qtZ8b8VTHt7GgDvz32fld+t5IVZL+DuXHrupSz4cAHJycnMnD6TaW9NIzMzkzM6nkHjJo0LnNelo0fz77FjaXX00fxj/Phc4xYvW8YH06dTvVo1ZsyezZLly/lg+nTSfv2Vdv360aZ5c156cx7t2p7MVUNHkJmZyZYtv7P0iyX8lPoT8976MHt7xap69Rq8NfNdHn/yUR569FHG33ord9//AFUqV+ad14NEtn7DBlLXrOGW22/njRnTqVq1KmcNHMTrs9+ka6eO/P7775xw/PHccN21nH/xJfzrrrt4ftIkvvr6a64YOYpBF17ExIkTqVq1KgsWLOCPP/6gTZs2dOrUKSG+j1IQJRWRiBx11FFcffXVXHPNNXTv3p0TTzxxhzK33347FStW5LLLLmPp0qUsXbqUIWf3AWB75nZq7V8737o3bdzIxo0baHFcGwB6nnEmwy8dHHNsDzz2NC+//jYAKatT+fq7laxZ+wsnHd+CWjWqA9C3RxdWfBt7p7bNmjfjoXseInV1Kh1P7UiDgxvsUOb9ue/z/tz3Of2U0wH4ffPvrPxuJZt/28wpXU+hYnJFANp3bl/gfNZv3MhvmzfT6uijgzi7d2fW3LnZ49u3bk31atUA+HDRIvqceipJSUnsX7MmbZo355OlS2nW9GiuHDGUjPQMunY+lSMbH8VBf2nAyh++57obRtHx5E60O+nkmJf91K7dAWhyVDNenRkk0P++/z4T7r03u0y1qlWZNX8+rVu1okaNGgD07tmDjxfMp2unjpQvX472bU8C4G+NGrFP+fKUK1eOwxs1IiUlBYDZs2fz+eefM3XqVCBIfCtWrFBSEdkbHHbYYXzyySfMnDmTMWPG0KFDh1zj33rrLaZMmcK8ecGvALs7jRs3ZuILr8Q1rnc/mM87//2Id195huSKFenYZxBb//gj5unLJpVl+/bgF2K3bv1zuu6nd6fJMU149613uejsi7jpjpuo/5f6uaZ1dy4cdiFn5rlHM+nhSbuwRLklV6xYZJnjW7VhxpTXePOd2Qy7+lIuHnIZ/fr0Z84b/2XOu+8w6enHmfHqdO698/6Y5lm+/D4AJJVJIjMzY6fiLlu2XPb3S8qUMcqXLx++LkNGZiYQrL/77ruPzp0779Q8SoPuqYhEZPXq1SQnJ3POOecwcuRIPvnkk+xxK1eu5LLLLmPKlClUDA+CjRo1Yu3atXy2aAEQ9CX29Zf/y7fuylWqUKVqNRZ9HDTVvPLSFJq3ah1TXBs3/cZ+VauQXLEiX379LfM/+RyAFkc34b8fLSRt3XrS09N58dXZ+U5/UP06fPr5FwBMe+3N7OGrVq6i/kH1GTBkAB06d+DLZV+SvG8ym3/bnF3mhPYn8NJzL7F5czBszU9rSFubRvPjmvP2rLfZumUrm3/bzJzZcwqMv1qVKuxbqRILFi8GyL5Pkp/Wxx7Li6+/TmZmJr+sW8cHCxdy7FFHsSrlB2rV2p8Bfx/I2f0H8PnSxaStS2P79u1079aDa0dez5Kli2NanwU56YQ2PPH009nv12/YwNFNm/Lh/PmkrVtHZmYm0195leNatoy5zs6dOzNhwgTS09MB+Oqrr7LXZaLSlYpIRJYsWcLIkSMpU6YM5cqVY8KECYwYMQKAJ554grS0NHr1Cm6m16lTh5kzZzJ16lSGXHQJmzZtIjMjgwGDL+KQRn/Lt/5xd93H2NEj2bplC/X+clCB917y6tTuBB556gWatj2Nw/7akJbHNAHgwNq1GHP1pbTtcTbVqlamSePc8zWCs+irLh7E2RdfzcRnptKlw0nZ42e9PIsZU2ZQrlw5au5fkwuvuJBq+1XjmJbHcFrb0zjp5JMY+Y+RfLPiG8469SwAkislc/sDt9O4SWO69uxKrw69qFGzBkc1O6rQZbj/5psZdsMNlClThjYtWlClcuV8y53WsSPzFy+mda9emBljR4ygdq1avD5zDg/859+UK1eOSsmVuO/uh0hNXc0VIy5n+/btAFx/zQ0xrc+CXHnZZYz+x42079KVpKQkhg8bSrfOnRk9ciR9zz4Hd6dD+/Z06dgx5jqHDBnC999/zzHHHIO7U6tWLaZPn75LccabuXtpx1Ck5s2bu375URLN8uXLOfzw2G9sF2RXv6dSkOJ+TyWnq8aMo9lRhzPwzN75ji/oeyq7qu6v+Q//bfNm9q1UCYC7HnmE1LVruX107I/57ur3VAqTKN9TyW9/NLNF7t48yriKoisVEcnlxtvvY8GnnzNm+KWlHUq2N959l7seeYSMjAzq16nDQ7feWtohSQGUVEQSzD/HjOLThfNzDRtw/oX07lf4FwrX/7qO8886Pfv9PgQ3kF+fPJEa1avFPP8bRw3lxlFDixFxdIaPHcvHn36aa9glAwZwzumnc0a3bnGf/7VjRjB/4ce5hl1w/sWc1e/suM97T6HmL5GdtCc3fxWlpJu/dpWav0qOnv4SEZHIKKmIiEhklFRERCQyulEvIjFpffeiSOt7ZkTRXwK89PrrmTV3LrWqV+fjV3bsecDdGTVuHLPnzSO5QgUmjBtHs8YF9yEm8acrFRFJWGf36sVLDz9c4PjZ8+bxzcqVfDZrFvfedBNXjR1bgtFJfpRURCRhtWnRgv2qFfw49Mx33uGsnj0xM1o2a8aGjRtJ/fnnEoxQ8lJSEZHd1uo1a6h3wJ+PC9c94ABWK6mUKiUVERGJjJKKiOy26tSuTUrqn7+W+WNqKnX2378UIxIlFRHZbXVt357nZszA3Zn/2WdUqVyZA5RUSpUeKRaRmHxw1bHZr+PVTUte5119Ne/Nn0/a+vX8rV07Rl9+OekZQZ9mg/v3p3PbtsyeN4+mnTuTXKECD44bVyJxScGUVEQkYT2e5/fo8zIz7rph134HRaKl5i8REYmMkoqIiEQmrknFzK4ysy/MbKmZPWdmFcysoZl9bGZfm9lkMysfzxhERKTkxC2pmFldYBjQ3N2PBJKA/sC/gLvd/RDgV2BwvGIQEZGSFe/mr7JARTMrCyQDPwEnA1PD8ZOAXnGOQURESkjckoq7/wjcCfxAkEw2AIuA9e6eERZLAermN72ZXWhmC81s4dq1a+MVpoiIRChujxSb2X5AT6AhsB6YAnSJdXp3fxh4GIKfE45HjCISuzWP9s9+nRxBfb9fNKXIMik//cRF117Lz2lpGDCoXz8uPffcXGXU/X1iief3VE4BvnP3tQBm9hLQBqhmZmXDq5V6wI9xjEFEdmNlk5K4ZdQomjVuzKbNmznpjDM4uXVr/nbIIdllcnZ/v2DxYq4aO5Y5kyeXYtR7t3jeU/kBOM7Mks3MgA7AMmAO0CcsMxCYEccYRGQ3dsD++2dfdVSuVIlGf/0rq9esyVVG3d8nlnjeU/mY4Ib8J8CScF4PA9cAw83sa6AGMDFeMYjInmPljz/y+fLlNG/aNNdwdX+fWOLaTYu7/wP4R57B3wJF/46oiEjot82bGTBsGLddey1V9t23tMORQugb9SKS0NLT0znniivod9pp9OjUaYfx6v4+sSipiEjCcncuGzOGRgcfzOWDBuVbRt3fJxb1UiwiMak95Pns1yXV9f1Hn3zC8y+/TOPDDqNN794A3HDllaT89BOg7u8TkZKKiCSs4489lo3LlxdaRt3fJxY1f4mISGSUVEREJDJKKiIiEhklFRERiYySioiIREZJRUREIqNHikUkJie/eGGk9U3sV3i3f1v/+IMuAwawbds2MjIy6Nm5M9cPHZqrzB/btnHRNdfw6bJlVK9WjSfuuouD6ub7E01SQnSlIiIJaZ/y5Xn18cf5YPp03p82jbfee4/5n32Wq8yTU6dSrWpVFr/xBpedey7/uPPOUopWsiipiEhCMjP2rVQJgPSMDDLS0wl+ReNPr4Xd3gP06tyZuR99hLt+0680KamISMLKzMykTe/e/PWEE2jfujUt8nR7/9OaNdQ78EAAypYtS5XKlVm3fn1phCohJRURSVhJSUm8P20ay+fMYdGSJSz76qvSDkmKoKQiIgmvWpUqnNiyJW+9916u4QfWrp3duWRGRgYbN22ierVqpRGihJRURCQh/bJuHes3bgRgy9atzPnwQw5t2DBXmW5ht/cA0994g7bHHbfDfRcpWXqkWERi8s4ZD2e/Lomu71PXruXi664jMzOT7du307tLF7q2b8/N//43xxx5JN1OPplz+/ThwmuuoWnnzuxXtSqPjx8f97ikcEoqIpKQjmzUiPdeemmH4WOGDct+XWGffXjynntKMiwpgpq/REQkMkoqIiISGSUVERGJjJKKiIhERklFREQio6QiIiKR0SPFIhKTj/pfGml9taf8J6ZymZmZtO3blwP3358pDz2Ua5y6vk88ulIRkYQ24amnOOzgg/Mdp67vE4+SiogkrB9TU3nj3XcZ2KdPvuPV9X3iUVIRkYR17a23MnbECMqUyf9Qpa7vE4+SiogkpNfnzKFm9eoc3bhxaYcixaAb9SKSkD7+9FNenzOHN+fNY+u2bWz67TeGjBrFo7ffnl0mq+v7ugccoK7vE4SuVEQkId04fDj/mzuXpW+/zePjx3NSq1a5Egqo6/tEpCsVEYnJcc8/mP26JLq+L4i6vk9sSioikvBObNmSE1u2BNT1faJT85eIiEQmrknFzKqZ2VQz+5+ZLTez482supm9aWYrwv/7xTMGEREpOfG+UrkXmOXufwOaAsuBa4G33f1Q4O3wvYiI7AHillTMrCpwEjARwN23uft6oCcwKSw2CegVrxhERKRkxfNKpSGwFnjczD41s0fNrBJQ291/CsukArXzm9jMLjSzhWa2cO3atXEMU0REohLPpFIWOAaY4O5HA5vJ09TlQSc9+XbU4+4Pu3tzd29eq1atOIYpIiJRiecjxSlAirt/HL6fSpBU1pjZge7+k5kdCPwcxxhEJCKPjP820vpOGXVoTOWO7NCBfStVIikpibJJSbw7dWqu8e7OqHHjmD1vHskVKjBh3DiaqWuXUhO3pOLuqWa2yswaufuXQAdgWfg3ELgt/D8jXjGIyJ7htUmTqLFf/g+Kzp43j29WruSzWbNYsHgxV40dy5zJk0s4QskXBDHiAAALc0lEQVQS7y8/DgWeMbPywLfAeQRNbi+Y2WBgJdAvzjGIyB5sZtj9vZnRslkzNmzcSOrPP3PA/vuXdmh7pbgmFXf/DGiez6gO8ZyviOw5zIxegwdjZpx35pmc1y/3eejqNWuod8AB2e/rHnAAq5VUSo26aRGRhPbGM89Qp3Zt1qal0XPwYA5r2JA2LVqUdlhSAHXTIiIJrU7t4FsHtWrUoPspp7BoyZIdxqekpma//zE1lTq6Sik1SioikrA2//47mzZvzn79zvvvc/ihuZ8a6xp2f+/uzP/sM6pUrqymr1Kk5i8RickFVx+c/bqkur7/OS2Ns4cOBSAjI4O+3bvT8cQTmfj88wAM7t+fzm3bMnvePJp27kxyhQo8OG5cicQm+VNSEZGE1bB+fT6YPn2H4YP7989+bWbcdcMNJRmWFELNXyIiEhklFRERiYySioiIRCameypm9ra7dyhqWCI7duSTcal3WuU74lLvWftViUu9AOOmxOdW2pIWI+JS74Y1k4outBOuf3pq0YVEpFgKPbqYWQUgGagZ/kKjhaOqAHXjHJuIiOxmijplvQi4EqgDLOLPpLIRuD+OcYmIyG6o0KTi7vcC95rZUHe/r4RiEpEEdOeof0Ra3xnjb46p3PqNGxn6f//HshUrMDMeuPlmWh19dPZ4dX2fWGJqXHf3+8ysNdAg5zTuHp8bFSIioWvGjeOUE07gqXvvZdu2bfy+dWuu8er6PrHEeqP+KeCvwGdAZjjYASUVEYmbDZs28cHChTx0660AlC9fnvLly+cqo67vE0usjwE1B44If/5XRKRErExJoUb16lwyejRLv/ySZkccwb9Gj6ZScnJ2GXV9n1hi/Z7KUuCAIkuJiEQoIzOTxcuWMbh/f9576SWSk5O565FHSjssKUSsSaUmsMzM3jCzl7P+4hmYiEjd2rWpW7s2LZo2BaBXp04sXrYsVxl1fZ9YYm3+ujGeQYiI5Kd2rVrUPfBAVnz3HYc2bMjcjz7ib4cckqtM1/btefjZZ+nTrRsLFi9W1/elLNanv96NdyAikthG3H5T9uuS6voe4I7rr2fIyJFsS0+nQf36PHjLLer6PoHF+vTXJoKnvQDKA+WAze4ev75ERESAJocfzrtTc3epo67vE1esVyqVs16bmQE9gePiFZSIiOyeit1LsQemA53jEI+IiOzGYm3+Oj3H2zIE31vZWkBxkb2GuxNcvIuUnkT6CmGsT3+dluN1BvA9QROYyF6rQoUKpKWlUaNGDSUWKTXuTlpaGhUqVCjtUIDY76mcF+9ARHY39erVIyUlhbVr1+5SPam//hZRRLll2sa41AvwS5n4/L7flt/jUi3p+2yOT8XA9sz41L3+j20xl61QoQL16tWLSxzFFWvzVz3gPqBNOOi/wBXunhKvwEQSXbly5WjYsOEu13PObvYDcgAXx+lH5Ha3H5AD/YhcXrGebjwOvEzwuyp1gFfCYSIiItliTSq13P1xd88I/54AasUxLhER2Q3FmlTSzOwcM0sK/84B0uIZmIiI7H5iTSrnA/2AVOAnoA8wKE4xiYjIbirWu2JjgYHu/iuAmVUH7iRINiIiIkDsVypNshIKgLuvA44upLyIiOyFYk0qZcxsv6w34ZVKfJ79ExGR3VasiWE88KGZTQnf9wVuiU9IIiKyu4r1G/VPmtlC4ORw0OnuvqywaUREZO8TcxNWmESUSEREpEDx6cAnh/B7LZ+a2avh+4Zm9rGZfW1mk82sfLxjEBGRkhH3pAJcASzP8f5fwN3ufgjwKzC4BGIQEZESENekEnZEeSrwaPjeCO7LZPWUNgnoFc8YRESk5MT7SuUeYBSwPXxfA1jv7hnh+xSgbn4TmtmFZrbQzBbuatfiIiJSMuKWVMysO/Czuy/amend/WF3b+7uzWvVUt+VIiK7g3h+gbEN0MPMugEVgCrAvUA1MysbXq3UA36MYwwiIlKC4nal4u7XuXs9d28A9AfecfezgTkEHVICDARmxCsGEREpWSXx9Fde1wDDzexrgnssE0shBhERiYMS6b/L3ecCc8PX3wItS2K+IiJSskrjSkVERPZQSioiIhIZJRUREYmMkoqIiERGSUVERCKjpCIiIpFRUhERkcgoqYiISGSUVEREJDJKKiIiEhklFRERiYySioiIREZJRUREIqOkIiIikVFSERGRyCipiIhIZJRUREQkMkoqIiISGSUVERGJjJKKiIhERklFREQio6QiIiKRUVIREZHIKKmIiEhklFRERCQySioiIhIZJRUREYmMkoqIiERGSUVERCKjpCIiIpFRUhERkcgoqYiISGSUVEREJDJKKiIiEhklFRERiUzckoqZ1TezOWa2zMy+MLMrwuHVzexNM1sR/t8vXjGIiEjJiueVSgZwtbsfARwHXGZmRwDXAm+7+6HA2+F7ERHZA8Qtqbj7T+7+Sfh6E7AcqAv0BCaFxSYBveIVg4iIlKwSuadiZg2Ao4GPgdru/lM4KhWoXcA0F5rZQjNbuHbt2pIIU0REdlHck4qZ7Qu8CFzp7htzjnN3Bzy/6dz9YXdv7u7Na9WqFe8wRUQkAnFNKmZWjiChPOPuL4WD15jZgeH4A4Gf4xmDiIiUnHg+/WXARGC5u9+VY9TLwMDw9UBgRrxiEBGRklU2jnW3AQYAS8zss3DYaOA24AUzGwysBPrFMQYRESlBcUsq7v4eYAWM7hCv+YqISOnRN+pFRCQySioiIhIZJRUREYmMkoqIiERGSUVERCKjpCIiIpFRUhERkcgoqYiISGSUVEREJDJKKiIiEhklFRERiYySioiIREZJRUREIqOkIiIikVFSERGRyCipiIhIZJRUREQkMkoqIiISGSUVERGJjJKKiIhERklFREQio6QiIiKRUVIREZHIKKmIiEhklFRERCQySioiIhIZJRUREYmMkoqIiERGSUVERCKjpCIiIpFRUhERkcgoqYiISGSUVEREJDJKKiIiEhklFRERiYySioiIRKZUkoqZdTGzL83sazO7tjRiEBGR6JV4UjGzJOABoCtwBHCWmR1R0nGIiEj0SuNKpSXwtbt/6+7bgOeBnqUQh4iIRMzcvWRnaNYH6OLuQ8L3A4BW7n55nnIXAheGbxsBX5ZooHu2msAvpR2ESD60b0brIHevVZIzLFuSMysOd38YeLi049gTmdlCd29e2nGI5KV9c/dXGs1fPwL1c7yvFw4TEZHdXGkklQXAoWbW0MzKA/2Bl0shDhERiViJN3+5e4aZXQ68ASQBj7n7FyUdx15OzYqSqLRv7uZK/Ea9iIjsufSNehERiYySioiIREZJZS+i7nEkUZnZY2b2s5ktLe1YZNcoqewl1D2OJLgngC6lHYTsOiWVvYe6x5GE5e7zgHWlHYfsOiWVvUddYFWO9ynhMBGRyCipiIhIZJRU9h7qHkdE4k5JZe+h7nFEJO6UVPYS7p4BZHWPsxx4Qd3jSKIws+eAD4FGZpZiZoNLOybZOeqmRUREIqMrFRERiYySioiIREZJRUREIqOkIiIikVFSERGRyCipiORgZr/Fse5BZnZ/vOoXSQRKKiIiEhklFZEimFktM3vRzBaEf23MrIyZfW9m1XKUW2FmtfMrX5rxi5QkJRWRot0L3O3uLYAzgEfdfTswA+gNYGatgJXuvia/8qUTtkjJK1vaAYjsBk4BjjCzrPdVzGxfYDJwA/A4QV9qk4soL7LHU1IRKVoZ4Dh335pzoJl9CBxiZrWAXsDNRZQviVhFSpWav0SKNhsYmvXGzJoBeNBx3jTgLmC5u6cVVl5kb6CkIpJbcthLbtbfcGAY0NzMPjezZcDFOcpPBs7hz6YviigvskdTL8UiIhIZXamIiEhklFRERCQySioiIhIZJRUREYmMkoqIiERGSUVERCKjpCIiIpH5fwZLEXAsvE+0AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# How does this play out at a Borough specific level? Let's look at Manhattan and Bronx\n",
    "bordf = df[df['Borough']=='M']\n",
    "sns.countplot(x = 'Level', hue = 'size_of_adjusted_gross_income', data = bordf)\n",
    "plt.title(\"Distribution of Accessibility Levels by Income in the Manhattan\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(x = 'size_of_adjusted_gross_income', hue = 'Level', data = bordf)\n",
    "plt.title(\"Distribution of Accessibility Levels by Income in the Manhattan\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The second plot in each Borough category shows that the availability distribution of playgrounds is almost identical accross all income groups, suggesting that these variables are not strongly correlated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Make a correlation plot to confirm. Convert the Level column to numeric\n",
    "df['Level'] = df.Level.astype('float32')\n",
    "sns.heatmap(df.corr())\n",
    "plt.title('Correlation Matrix Plot showing different features and their correlation with each other')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Accessible                                     0.076418\n",
       "Adaptive_Swing                                -0.572829\n",
       "Level                                          1.000000\n",
       "num_of_returns                                 0.004581\n",
       "num_of_single_returns                         -0.000939\n",
       "size_of_adjusted_gross_income                  0.000000\n",
       "num_of_head_of_household_returns               0.014223\n",
       "num_of_exemptions                              0.009715\n",
       "num_of_joint_returns                           0.005316\n",
       "num_of_dependents                              0.017000\n",
       "normalized_num_of_joint_returns                0.010519\n",
       "normalized_num_of_head_of_household_returns    0.029707\n",
       "normalized_num_of_single_returns              -0.033173\n",
       "Name: Level, dtype: float64"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.corr().Level"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Preprocessing\n",
    "\n",
    "For further preparation of the data for ML, we need to drop some more columns:\n",
    "1) We have subsumed the accessibility column under level, so we can drop that. Remember that anything that was not wheelchair accessible is classifed as Level 0 <br/>\n",
    "\n",
    "2) The adaptive swing column is correlated to Level 0 -- Not accessible, and can help pick that out. <br/> \n",
    "\n",
    "3) We will only keep the normalized columns for joint, HOH and Single returns and drop the unnormalized ones. <br/>\n",
    "\n",
    "4) We will keep the total number of returns column <br/> \n",
    "\n",
    "5) SInce the total returns is a pretty large number, we will standardize the columns as well <br/>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Machine Learning modeling\n",
    "\n",
    "Here we will try to predict the level of playground based on the data available.\n",
    "For any ML model, we need to provide the label column first. Let's drop some of the fields above and prepare the data for ML training in SageMaker"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_df = df.drop(columns =['Accessible', 'num_of_returns', \n",
    "                             'num_of_dependents','num_of_joint_returns', 'num_of_head_of_household_returns',\n",
    "                             'num_of_single_returns'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4590, 8)"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_df = pd.get_dummies(final_df)\n",
    "df1 = final_df.Level\n",
    "df2 = final_df.drop(columns = 'Level')\n",
    "final_df = pd.concat([df1, df2], axis = 1)\n",
    "final_df.dropna(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4564, 13)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7ff4dd74de80>"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "X_train, X_test = train_test_split(final_df, test_size = 0.2, random_state = 42)\n",
    "sns.countplot(X_test.Level)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "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>Level</th>\n",
       "      <th>Adaptive_Swing</th>\n",
       "      <th>size_of_adjusted_gross_income</th>\n",
       "      <th>num_of_exemptions</th>\n",
       "      <th>normalized_num_of_joint_returns</th>\n",
       "      <th>normalized_num_of_head_of_household_returns</th>\n",
       "      <th>normalized_num_of_single_returns</th>\n",
       "      <th>Borough_B</th>\n",
       "      <th>Borough_M</th>\n",
       "      <th>Borough_N</th>\n",
       "      <th>Borough_Q</th>\n",
       "      <th>Borough_R</th>\n",
       "      <th>Borough_X</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3848</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1620.0</td>\n",
       "      <td>0.111940</td>\n",
       "      <td>0.029851</td>\n",
       "      <td>0.820896</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1207</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>19010.0</td>\n",
       "      <td>0.076699</td>\n",
       "      <td>0.344660</td>\n",
       "      <td>0.570874</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5272</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>40.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>241</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>7640.0</td>\n",
       "      <td>0.113379</td>\n",
       "      <td>0.371882</td>\n",
       "      <td>0.485261</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3510.0</td>\n",
       "      <td>0.466216</td>\n",
       "      <td>0.297297</td>\n",
       "      <td>0.229730</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Level  Adaptive_Swing  size_of_adjusted_gross_income  num_of_exemptions  \\\n",
       "3848    1.0             0.0                            4.0             1620.0   \n",
       "1207    1.0             0.0                            1.0            19010.0   \n",
       "5272    0.0             1.0                            6.0               40.0   \n",
       "241     1.0             0.0                            3.0             7640.0   \n",
       "993     1.0             0.0                            5.0             3510.0   \n",
       "\n",
       "      normalized_num_of_joint_returns  \\\n",
       "3848                         0.111940   \n",
       "1207                         0.076699   \n",
       "5272                         0.000000   \n",
       "241                          0.113379   \n",
       "993                          0.466216   \n",
       "\n",
       "      normalized_num_of_head_of_household_returns  \\\n",
       "3848                                     0.029851   \n",
       "1207                                     0.344660   \n",
       "5272                                     0.000000   \n",
       "241                                      0.371882   \n",
       "993                                      0.297297   \n",
       "\n",
       "      normalized_num_of_single_returns  Borough_B  Borough_M  Borough_N  \\\n",
       "3848                          0.820896          0          1          0   \n",
       "1207                          0.570874          0          0          0   \n",
       "5272                          0.000000          0          0          0   \n",
       "241                           0.485261          1          0          0   \n",
       "993                           0.229730          0          0          0   \n",
       "\n",
       "      Borough_Q  Borough_R  Borough_X  \n",
       "3848          0          0          0  \n",
       "1207          0          0          1  \n",
       "5272          0          0          1  \n",
       "241           0          0          0  \n",
       "993           1          0          0  "
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Level                                          False\n",
       "Adaptive_Swing                                 False\n",
       "size_of_adjusted_gross_income                  False\n",
       "num_of_exemptions                              False\n",
       "normalized_num_of_joint_returns                False\n",
       "normalized_num_of_head_of_household_returns    False\n",
       "normalized_num_of_single_returns               False\n",
       "Borough_B                                      False\n",
       "Borough_M                                      False\n",
       "Borough_N                                      False\n",
       "Borough_Q                                      False\n",
       "Borough_R                                      False\n",
       "Borough_X                                      False\n",
       "dtype: bool"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.isnull().any()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Upload Training data to S3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "role = get_execution_role()\n",
    "session = sagemaker.Session()\n",
    "bucket = session.default_bucket() # you can replace with this your own bucket name\n",
    "prefix = 'sagemaker/accessibility'\n",
    "\n",
    "train_file = 'train.csv'\n",
    "X_train.to_csv(train_file,index=False,header=False)\n",
    "boto3.Session().resource('s3').Bucket(bucket).Object(os.path.join(prefix, 'train/train.csv')).upload_file('train.csv')\n",
    "train_data = sagemaker.session.s3_input('s3://{}/{}/train'.format(bucket, prefix), \n",
    "                                        content_type='csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "container = get_image_uri(boto3.Session().region_name, 'xgboost', '0.90-1')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training artifacts will be uploaded to: s3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/output\n"
     ]
    }
   ],
   "source": [
    "output_location = 's3://{}/{}/output'.format(bucket, prefix)\n",
    "print('training artifacts will be uploaded to: {}'.format(output_location))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2019-12-23 19:06:32 Starting - Starting the training job...\n",
      "2019-12-23 19:06:34 Starting - Launching requested ML instances......\n",
      "2019-12-23 19:07:39 Starting - Preparing the instances for training......\n",
      "2019-12-23 19:08:59 Downloading - Downloading input data\n",
      "2019-12-23 19:08:59 Training - Downloading the training image...\n",
      "2019-12-23 19:09:32 Uploading - Uploading generated training model.\u001b[31mINFO:sagemaker-containers:Imported framework sagemaker_xgboost_container.training\u001b[0m\n",
      "\u001b[31mINFO:sagemaker-containers:Failed to parse hyperparameter objective value binary:logistic to Json.\u001b[0m\n",
      "\u001b[31mReturning the value itself\u001b[0m\n",
      "\u001b[31mINFO:sagemaker-containers:No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[31mINFO:sagemaker_xgboost_container.training:Running XGBoost Sagemaker in algorithm mode\u001b[0m\n",
      "\u001b[31mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[31mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[31mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[31m[19:09:29] 3651x12 matrix with 43812 entries loaded from /opt/ml/input/data/train?format=csv&label_column=0&delimiter=,\u001b[0m\n",
      "\u001b[31mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[31m[19:09:29] 3651x12 matrix with 43812 entries loaded from /opt/ml/input/data/validation?format=csv&label_column=0&delimiter=,\u001b[0m\n",
      "\u001b[31mINFO:root:Single node training.\u001b[0m\n",
      "\u001b[31mINFO:root:Train matrix has 3651 rows\u001b[0m\n",
      "\u001b[31mINFO:root:Validation matrix has 3651 rows\u001b[0m\n",
      "\u001b[31m[0]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[1]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[2]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[3]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[4]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[5]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[6]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[7]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[8]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[9]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[10]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[11]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[12]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[13]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[14]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[15]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[16]#011train-error:0.178307#011validation-error:0.178307\u001b[0m\n",
      "\u001b[31m[17]#011train-error:0.177486#011validation-error:0.177486\u001b[0m\n",
      "\u001b[31m[18]#011train-error:0.177486#011validation-error:0.177486\u001b[0m\n",
      "\u001b[31m[19]#011train-error:0.177486#011validation-error:0.177486\u001b[0m\n",
      "\u001b[31m[20]#011train-error:0.177486#011validation-error:0.177486\u001b[0m\n",
      "\u001b[31m[21]#011train-error:0.177486#011validation-error:0.177486\u001b[0m\n",
      "\u001b[31m[22]#011train-error:0.177486#011validation-error:0.177486\u001b[0m\n",
      "\u001b[31m[23]#011train-error:0.177486#011validation-error:0.177486\u001b[0m\n",
      "\u001b[31m[24]#011train-error:0.177486#011validation-error:0.177486\u001b[0m\n",
      "\u001b[31m[25]#011train-error:0.177486#011validation-error:0.177486\u001b[0m\n",
      "\u001b[31m[26]#011train-error:0.177486#011validation-error:0.177486\u001b[0m\n",
      "\u001b[31m[27]#011train-error:0.177486#011validation-error:0.177486\u001b[0m\n",
      "\u001b[31m[28]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[29]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[30]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[31]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[32]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[33]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[34]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[35]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[36]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[37]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[38]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[39]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[40]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[41]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[42]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[43]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[44]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[45]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[46]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[47]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[48]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[49]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[50]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[51]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[52]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[53]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[54]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[55]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[56]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[57]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[58]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[59]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[60]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[61]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[62]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[63]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[64]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[65]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[66]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[67]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[68]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[69]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[70]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[71]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[72]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[73]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[74]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[75]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[76]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[77]#011train-error:0.176664#011validation-error:0.176664\u001b[0m\n",
      "\u001b[31m[78]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[79]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[80]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[81]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[82]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[83]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[84]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[85]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[86]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[87]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[88]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[89]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[90]#011train-error:0.175568#011validation-error:0.175568\u001b[0m\n",
      "\u001b[31m[91]#011train-error:0.175568#011validation-error:0.175568\u001b[0m\n",
      "\u001b[31m[92]#011train-error:0.175568#011validation-error:0.175568\u001b[0m\n",
      "\u001b[31m[93]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[94]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[95]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[96]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[97]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[98]#011train-error:0.176116#011validation-error:0.176116\u001b[0m\n",
      "\u001b[31m[99]#011train-error:0.175568#011validation-error:0.175568\u001b[0m\n",
      "\n",
      "2019-12-23 19:09:39 Completed - Training job completed\n",
      "Training seconds: 58\n",
      "Billable seconds: 58\n"
     ]
    }
   ],
   "source": [
    "sess = sagemaker.Session()\n",
    "\n",
    "xgb = sagemaker.estimator.Estimator(container,\n",
    "                                       role, \n",
    "                                       train_instance_count=1, \n",
    "                                       train_instance_type='ml.c4.xlarge',\n",
    "                                       output_path=output_location,\n",
    "                                       sagemaker_session=sess)\n",
    "\n",
    "xgb.set_hyperparameters(max_depth=5,\n",
    "                        eta=0.2,\n",
    "                        gamma=4,\n",
    "                        min_child_weight=6,\n",
    "                        subsample=0.8,\n",
    "                        silent=0,\n",
    "                        objective='binary:logistic',\n",
    "                        num_round=100)\n",
    "\n",
    "xgb.fit({'train': train_data, 'validation': train_data})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Using SageMaker Autopilot\n",
    "\n",
    "As you can see, training an ML model, feature engineering, preprocessing can take quite some time and be a fairly involved process. Let's see how we can use SageMaker Autopilot, which is SageMaker's built in AutoML feature to take the raw dataset, and train a bunch of ML models, and let AutoPilot take care of the preprocessing, feature engineering and model training all within a few simple API calls. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Lets start by looking at autodf -- we will drop the accessibility column, since we have already incorporated that into Level. Note that to go from autodf to the final training dataset for XGBoost above we had to implement a sequence of feature engineering steps:\n",
    "\n",
    "1) We converted Borough columns to one hot encoding\n",
    "\n",
    "2) We dropped all the non normalized columns which are strongly correlated with the feature engineered normalized ones.\n",
    "\n",
    "3) We had to split the data randomly into training and test datasets which we uploaded separately to S3.\n",
    "\n",
    "4) We had to save the csv files locally without headers and indexes. \n",
    "\n",
    "5) While this is not required for XGBoost, we removed missing values. Linear learner algorithm for example doesn't accept missing values. \n",
    "\n",
    "Autopilot takes care of all the feature engineering steps for us. Simply feed in the initial dataframe as an input, save it as csv, point Autopilot to the dependent variable and it will take care of the rest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "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>Accessible</th>\n",
       "      <th>Adaptive_Swing</th>\n",
       "      <th>Level</th>\n",
       "      <th>num_of_returns</th>\n",
       "      <th>num_of_single_returns</th>\n",
       "      <th>size_of_adjusted_gross_income</th>\n",
       "      <th>num_of_head_of_household_returns</th>\n",
       "      <th>num_of_exemptions</th>\n",
       "      <th>num_of_joint_returns</th>\n",
       "      <th>num_of_dependents</th>\n",
       "      <th>Borough</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>10120.0</td>\n",
       "      <td>6260.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2310.0</td>\n",
       "      <td>16270.0</td>\n",
       "      <td>1380.0</td>\n",
       "      <td>5480.0</td>\n",
       "      <td>B</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>4890.0</td>\n",
       "      <td>2310.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1510.0</td>\n",
       "      <td>9100.0</td>\n",
       "      <td>960.0</td>\n",
       "      <td>3260.0</td>\n",
       "      <td>B</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2350.0</td>\n",
       "      <td>1050.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>580.0</td>\n",
       "      <td>4320.0</td>\n",
       "      <td>630.0</td>\n",
       "      <td>1350.0</td>\n",
       "      <td>B</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1280.0</td>\n",
       "      <td>470.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>260.0</td>\n",
       "      <td>2490.0</td>\n",
       "      <td>520.0</td>\n",
       "      <td>700.0</td>\n",
       "      <td>B</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1350.0</td>\n",
       "      <td>330.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>180.0</td>\n",
       "      <td>3100.0</td>\n",
       "      <td>810.0</td>\n",
       "      <td>950.0</td>\n",
       "      <td>B</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Accessible  Adaptive_Swing  Level  num_of_returns  num_of_single_returns  \\\n",
       "0         1.0             0.0      1         10120.0                 6260.0   \n",
       "1         1.0             0.0      1          4890.0                 2310.0   \n",
       "2         1.0             0.0      1          2350.0                 1050.0   \n",
       "3         1.0             0.0      1          1280.0                  470.0   \n",
       "4         1.0             0.0      1          1350.0                  330.0   \n",
       "\n",
       "   size_of_adjusted_gross_income  num_of_head_of_household_returns  \\\n",
       "0                            1.0                            2310.0   \n",
       "1                            2.0                            1510.0   \n",
       "2                            3.0                             580.0   \n",
       "3                            4.0                             260.0   \n",
       "4                            5.0                             180.0   \n",
       "\n",
       "   num_of_exemptions  num_of_joint_returns  num_of_dependents Borough  \n",
       "0            16270.0                1380.0             5480.0       B  \n",
       "1             9100.0                 960.0             3260.0       B  \n",
       "2             4320.0                 630.0             1350.0       B  \n",
       "3             2490.0                 520.0              700.0       B  \n",
       "4             3100.0                 810.0              950.0       B  "
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#look at the autodf dataframe again\n",
    "autodf.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Accessible                           True\n",
       "Adaptive_Swing                       True\n",
       "Level                               False\n",
       "num_of_returns                       True\n",
       "num_of_single_returns                True\n",
       "size_of_adjusted_gross_income        True\n",
       "num_of_head_of_household_returns     True\n",
       "num_of_exemptions                    True\n",
       "num_of_joint_returns                 True\n",
       "num_of_dependents                    True\n",
       "Borough                             False\n",
       "dtype: bool"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# note that we have missing values in almost every column. We will drop those as AutoML doesn't support missing values\n",
    "#for now.\n",
    "autodf.isnull().any()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4590, 11)\n",
      "Shape of test data = (459, 11)\n",
      "s3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/input/automl-train.csv\n"
     ]
    }
   ],
   "source": [
    "#Load the autodf into s3 directly\n",
    "autodf.dropna(inplace=True)\n",
    "print(autodf.shape)\n",
    "autodf = autodf.sample(frac=1)\n",
    "\n",
    "#Let's take out a random sample for testing and set it aside\n",
    "test_sample = autodf[TRAIN_LENGTH:]\n",
    "print(\"Shape of test data = {}\".format(test_sample.shape))\n",
    "test_sample.drop(columns=['Accessible']).to_csv('automl-test.csv', index=False)\n",
    "\n",
    "\n",
    "autodf[:TRAIN_LENGTH].drop(columns=['Accessible']).to_csv('automl-train.csv', index= False)\n",
    "autotrainpath = session.upload_data(path=\"automl-train.csv\", key_prefix=prefix + \"/input\")\n",
    "print(autotrainpath)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[{'DataSource': {'S3DataSource': {'S3DataType': 'S3Prefix', 'S3Uri': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/input'}}, 'TargetAttributeName': 'Level'}]\n",
      "{'S3OutputPath': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/automloutput'}\n"
     ]
    }
   ],
   "source": [
    "# configure the AutoML job: specify the target attribute name and the location of the input file\n",
    "input_data_config = [{\n",
    "      'DataSource': {\n",
    "        'S3DataSource': {\n",
    "          'S3DataType': 'S3Prefix',\n",
    "          'S3Uri': 's3://{}/{}/input'.format(bucket,prefix)\n",
    "        }\n",
    "      },\n",
    "      'TargetAttributeName': 'Level'\n",
    "    \n",
    "    }\n",
    "  ]\n",
    "\n",
    "output_data_config = {\n",
    "    'S3OutputPath': 's3://{}/{}/automloutput'.format(bucket,prefix)\n",
    "  }\n",
    "print(input_data_config)\n",
    "print(output_data_config)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "from time import gmtime, strftime, sleep\n",
    "timestamp_suffix = strftime('%d-%H-%M-%S', gmtime())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AutoMLJobName: automl-dm-23-15-11-31\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'AutoMLJobArn': 'arn:aws:sagemaker:us-east-1:389535300735:automl-job/automl-dm-23-15-11-31',\n",
       " 'ResponseMetadata': {'RequestId': 'df06c6c1-7c0e-4cd8-81d3-40ca486679ff',\n",
       "  'HTTPStatusCode': 200,\n",
       "  'HTTPHeaders': {'x-amzn-requestid': 'df06c6c1-7c0e-4cd8-81d3-40ca486679ff',\n",
       "   'content-type': 'application/x-amz-json-1.1',\n",
       "   'content-length': '92',\n",
       "   'date': 'Mon, 23 Dec 2019 15:11:41 GMT'},\n",
       "  'RetryAttempts': 0}}"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "region = boto3.Session().region_name\n",
    "sm = boto3.Session().client(service_name='sagemaker',region_name=region)\n",
    "auto_ml_job_name = 'automl-dm-' + timestamp_suffix\n",
    "print('AutoMLJobName: ' + auto_ml_job_name)\n",
    "\n",
    "sm.create_auto_ml_job(AutoMLJobName=auto_ml_job_name,\n",
    "                      InputDataConfig=input_data_config,\n",
    "                      OutputDataConfig=output_data_config,\n",
    "                      RoleArn=role,\n",
    "                     ProblemType='BinaryClassification',\n",
    "                     AutoMLJobObjective = {'MetricName':'Accuracy'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "JobStatus - Secondary Status\n",
      "------------------------------\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - AnalyzingData\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - FeatureEngineering\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "InProgress - ModelTuning\n",
      "Completed - MaxCandidatesReached\n",
      "Time Taken for job = 7165.622578859329\n"
     ]
    }
   ],
   "source": [
    "# Kick off the Autopilot job\n",
    "import time\n",
    "start = time.time()\n",
    "print ('JobStatus - Secondary Status')\n",
    "print('------------------------------')\n",
    "\n",
    "\n",
    "describe_response = sm.describe_auto_ml_job(AutoMLJobName=auto_ml_job_name)\n",
    "print (describe_response['AutoMLJobStatus'] + \" - \" + describe_response['AutoMLJobSecondaryStatus'])\n",
    "job_run_status = describe_response['AutoMLJobStatus']\n",
    "    \n",
    "while job_run_status not in ('Failed', 'Completed', 'Stopped'):\n",
    "    describe_response = sm.describe_auto_ml_job(AutoMLJobName=auto_ml_job_name)\n",
    "    job_run_status = describe_response['AutoMLJobStatus']\n",
    "    \n",
    "    print (describe_response['AutoMLJobStatus'] + \" - \" + describe_response['AutoMLJobSecondaryStatus'])\n",
    "    sleep(30)\n",
    "end = time.time()\n",
    "print(\"Time Taken for job = {}\".format(end - start))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'AutoMLJobName': 'automl-dm-23-15-11-31',\n",
       " 'AutoMLJobArn': 'arn:aws:sagemaker:us-east-1:389535300735:automl-job/automl-dm-23-15-11-31',\n",
       " 'InputDataConfig': [{'DataSource': {'S3DataSource': {'S3DataType': 'S3Prefix',\n",
       "     'S3Uri': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/input'}},\n",
       "   'TargetAttributeName': 'Level'}],\n",
       " 'OutputDataConfig': {'S3OutputPath': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/automloutput'},\n",
       " 'RoleArn': 'arn:aws:iam::389535300735:role/service-role/AmazonSageMaker-ExecutionRole-20190823T110499',\n",
       " 'AutoMLJobObjective': {'MetricName': 'Accuracy'},\n",
       " 'ProblemType': 'BinaryClassification',\n",
       " 'CreationTime': datetime.datetime(2019, 12, 23, 15, 11, 41, 698000, tzinfo=tzlocal()),\n",
       " 'EndTime': datetime.datetime(2019, 12, 23, 17, 11, 31, 899000, tzinfo=tzlocal()),\n",
       " 'LastModifiedTime': datetime.datetime(2019, 12, 23, 17, 11, 31, 936000, tzinfo=tzlocal()),\n",
       " 'BestCandidate': {'CandidateName': 'tuning-job-1-e4229d391ccb4660b4-086-00948424',\n",
       "  'FinalAutoMLJobObjectiveMetric': {'MetricName': 'validation:accuracy',\n",
       "   'Value': 0.8189550042152405},\n",
       "  'ObjectiveStatus': 'Succeeded',\n",
       "  'CandidateSteps': [{'CandidateStepType': 'AWS::SageMaker::ProcessingJob',\n",
       "    'CandidateStepArn': 'arn:aws:sagemaker:us-east-1:389535300735:processing-job/db-1-694f3608008c4db3b0ef60965eb5c9f476d3af96506144e49ff7d62b16',\n",
       "    'CandidateStepName': 'db-1-694f3608008c4db3b0ef60965eb5c9f476d3af96506144e49ff7d62b16'},\n",
       "   {'CandidateStepType': 'AWS::SageMaker::TrainingJob',\n",
       "    'CandidateStepArn': 'arn:aws:sagemaker:us-east-1:389535300735:training-job/automl-dm--dpp0-1-779b98a64b5c402bb77c98c96131c3e96841b2d538bc4',\n",
       "    'CandidateStepName': 'automl-dm--dpp0-1-779b98a64b5c402bb77c98c96131c3e96841b2d538bc4'},\n",
       "   {'CandidateStepType': 'AWS::SageMaker::TransformJob',\n",
       "    'CandidateStepArn': 'arn:aws:sagemaker:us-east-1:389535300735:transform-job/automl-dm--dpp0-csv-1-b82f119579a84577a7b3c78d83c85030438f769d1',\n",
       "    'CandidateStepName': 'automl-dm--dpp0-csv-1-b82f119579a84577a7b3c78d83c85030438f769d1'},\n",
       "   {'CandidateStepType': 'AWS::SageMaker::TrainingJob',\n",
       "    'CandidateStepArn': 'arn:aws:sagemaker:us-east-1:389535300735:training-job/tuning-job-1-e4229d391ccb4660b4-086-00948424',\n",
       "    'CandidateStepName': 'tuning-job-1-e4229d391ccb4660b4-086-00948424'}],\n",
       "  'CandidateStatus': 'Completed',\n",
       "  'InferenceContainers': [{'Image': '683313688378.dkr.ecr.us-east-1.amazonaws.com/sagemaker-sklearn-automl:0.1.0-cpu-py3',\n",
       "    'ModelDataUrl': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/automloutput/automl-dm-23-15-11-31/data-processor-models/automl-dm--dpp0-1-779b98a64b5c402bb77c98c96131c3e96841b2d538bc4/output/model.tar.gz',\n",
       "    'Environment': {'AUTOML_TRANSFORM_MODE': 'feature-transform',\n",
       "     'SAGEMAKER_DEFAULT_INVOCATIONS_ACCEPT': 'application/x-recordio-protobuf',\n",
       "     'SAGEMAKER_PROGRAM': 'sagemaker_serve',\n",
       "     'SAGEMAKER_SUBMIT_DIRECTORY': '/opt/ml/model/sagemaker_serve.py'}},\n",
       "   {'Image': '683313688378.dkr.ecr.us-east-1.amazonaws.com/sagemaker-xgboost:0.90-1-cpu-py3',\n",
       "    'ModelDataUrl': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/automloutput/automl-dm-23-15-11-31/tuning/automl-dm--dpp0-xgb/tuning-job-1-e4229d391ccb4660b4-086-00948424/output/model.tar.gz',\n",
       "    'Environment': {'MAX_CONTENT_LENGTH': '20971520',\n",
       "     'SAGEMAKER_DEFAULT_INVOCATIONS_ACCEPT': 'text/csv'}},\n",
       "   {'Image': '683313688378.dkr.ecr.us-east-1.amazonaws.com/sagemaker-sklearn-automl:0.1.0-cpu-py3',\n",
       "    'ModelDataUrl': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/automloutput/automl-dm-23-15-11-31/data-processor-models/automl-dm--dpp0-1-779b98a64b5c402bb77c98c96131c3e96841b2d538bc4/output/model.tar.gz',\n",
       "    'Environment': {'AUTOML_TRANSFORM_MODE': 'inverse-label-transform',\n",
       "     'SAGEMAKER_DEFAULT_INVOCATIONS_ACCEPT': 'text/csv',\n",
       "     'SAGEMAKER_PROGRAM': 'sagemaker_serve',\n",
       "     'SAGEMAKER_SUBMIT_DIRECTORY': '/opt/ml/model/sagemaker_serve.py'}}],\n",
       "  'CreationTime': datetime.datetime(2019, 12, 23, 17, 11, 29, 969000, tzinfo=tzlocal()),\n",
       "  'LastModifiedTime': datetime.datetime(2019, 12, 23, 17, 11, 29, 969000, tzinfo=tzlocal())},\n",
       " 'AutoMLJobStatus': 'Completed',\n",
       " 'AutoMLJobSecondaryStatus': 'MaxCandidatesReached',\n",
       " 'GenerateCandidateDefinitionsOnly': False,\n",
       " 'AutoMLJobArtifacts': {'CandidateDefinitionNotebookLocation': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/automloutput/automl-dm-23-15-11-31/sagemaker-automl-candidates/pr-1-ee4c8e4cb21f4a568cecb742bf02f24a6fa7e61f18ab4ac79ddf372968/notebooks/SageMakerAutopilotCandidateDefinitionNotebook.ipynb',\n",
       "  'DataExplorationNotebookLocation': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/automloutput/automl-dm-23-15-11-31/sagemaker-automl-candidates/pr-1-ee4c8e4cb21f4a568cecb742bf02f24a6fa7e61f18ab4ac79ddf372968/notebooks/SageMakerAutopilotDataExplorationNotebook.ipynb'},\n",
       " 'ResolvedAttributes': {'AutoMLJobObjective': {'MetricName': 'Accuracy'},\n",
       "  'ProblemType': 'BinaryClassification',\n",
       "  'CompletionCriteria': {'MaxCandidates': 250,\n",
       "   'MaxRuntimePerTrainingJobInSeconds': 86400,\n",
       "   'MaxAutoMLJobRuntimeInSeconds': 2592000}},\n",
       " 'ResponseMetadata': {'RequestId': '4657086e-3543-4629-9210-6ba3bc7ae0b7',\n",
       "  'HTTPStatusCode': 200,\n",
       "  'HTTPHeaders': {'x-amzn-requestid': '4657086e-3543-4629-9210-6ba3bc7ae0b7',\n",
       "   'content-type': 'application/x-amz-json-1.1',\n",
       "   'content-length': '4564',\n",
       "   'date': 'Mon, 23 Dec 2019 17:11:54 GMT'},\n",
       "  'RetryAttempts': 0}}"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "describe_response"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'CandidateName': 'tuning-job-1-e4229d391ccb4660b4-086-00948424', 'FinalAutoMLJobObjectiveMetric': {'MetricName': 'validation:accuracy', 'Value': 0.8189550042152405}, 'ObjectiveStatus': 'Succeeded', 'CandidateSteps': [{'CandidateStepType': 'AWS::SageMaker::ProcessingJob', 'CandidateStepArn': 'arn:aws:sagemaker:us-east-1:389535300735:processing-job/db-1-694f3608008c4db3b0ef60965eb5c9f476d3af96506144e49ff7d62b16', 'CandidateStepName': 'db-1-694f3608008c4db3b0ef60965eb5c9f476d3af96506144e49ff7d62b16'}, {'CandidateStepType': 'AWS::SageMaker::TrainingJob', 'CandidateStepArn': 'arn:aws:sagemaker:us-east-1:389535300735:training-job/automl-dm--dpp0-1-779b98a64b5c402bb77c98c96131c3e96841b2d538bc4', 'CandidateStepName': 'automl-dm--dpp0-1-779b98a64b5c402bb77c98c96131c3e96841b2d538bc4'}, {'CandidateStepType': 'AWS::SageMaker::TransformJob', 'CandidateStepArn': 'arn:aws:sagemaker:us-east-1:389535300735:transform-job/automl-dm--dpp0-csv-1-b82f119579a84577a7b3c78d83c85030438f769d1', 'CandidateStepName': 'automl-dm--dpp0-csv-1-b82f119579a84577a7b3c78d83c85030438f769d1'}, {'CandidateStepType': 'AWS::SageMaker::TrainingJob', 'CandidateStepArn': 'arn:aws:sagemaker:us-east-1:389535300735:training-job/tuning-job-1-e4229d391ccb4660b4-086-00948424', 'CandidateStepName': 'tuning-job-1-e4229d391ccb4660b4-086-00948424'}], 'CandidateStatus': 'Completed', 'InferenceContainers': [{'Image': '683313688378.dkr.ecr.us-east-1.amazonaws.com/sagemaker-sklearn-automl:0.1.0-cpu-py3', 'ModelDataUrl': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/automloutput/automl-dm-23-15-11-31/data-processor-models/automl-dm--dpp0-1-779b98a64b5c402bb77c98c96131c3e96841b2d538bc4/output/model.tar.gz', 'Environment': {'AUTOML_TRANSFORM_MODE': 'feature-transform', 'SAGEMAKER_DEFAULT_INVOCATIONS_ACCEPT': 'application/x-recordio-protobuf', 'SAGEMAKER_PROGRAM': 'sagemaker_serve', 'SAGEMAKER_SUBMIT_DIRECTORY': '/opt/ml/model/sagemaker_serve.py'}}, {'Image': '683313688378.dkr.ecr.us-east-1.amazonaws.com/sagemaker-xgboost:0.90-1-cpu-py3', 'ModelDataUrl': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/automloutput/automl-dm-23-15-11-31/tuning/automl-dm--dpp0-xgb/tuning-job-1-e4229d391ccb4660b4-086-00948424/output/model.tar.gz', 'Environment': {'MAX_CONTENT_LENGTH': '20971520', 'SAGEMAKER_DEFAULT_INVOCATIONS_ACCEPT': 'text/csv'}}, {'Image': '683313688378.dkr.ecr.us-east-1.amazonaws.com/sagemaker-sklearn-automl:0.1.0-cpu-py3', 'ModelDataUrl': 's3://sagemaker-us-east-1-389535300735/sagemaker/accessibility/automloutput/automl-dm-23-15-11-31/data-processor-models/automl-dm--dpp0-1-779b98a64b5c402bb77c98c96131c3e96841b2d538bc4/output/model.tar.gz', 'Environment': {'AUTOML_TRANSFORM_MODE': 'inverse-label-transform', 'SAGEMAKER_DEFAULT_INVOCATIONS_ACCEPT': 'text/csv', 'SAGEMAKER_PROGRAM': 'sagemaker_serve', 'SAGEMAKER_SUBMIT_DIRECTORY': '/opt/ml/model/sagemaker_serve.py'}}], 'CreationTime': datetime.datetime(2019, 12, 23, 17, 11, 29, 969000, tzinfo=tzlocal()), 'LastModifiedTime': datetime.datetime(2019, 12, 23, 17, 11, 29, 969000, tzinfo=tzlocal())}\n",
      "\n",
      "\n",
      "CandidateName: tuning-job-1-e4229d391ccb4660b4-086-00948424\n",
      "FinalAutoMLJobObjectiveMetricName: validation:accuracy\n",
      "FinalAutoMLJobObjectiveMetricValue: 0.8189550042152405\n"
     ]
    }
   ],
   "source": [
    "# lets look at the best job\n",
    "best_candidate = sm.describe_auto_ml_job(AutoMLJobName=auto_ml_job_name)['BestCandidate']\n",
    "best_candidate_name = best_candidate['CandidateName']\n",
    "print(best_candidate)\n",
    "print('\\n')\n",
    "print(\"CandidateName: \" + best_candidate_name)\n",
    "print(\"FinalAutoMLJobObjectiveMetricName: \" + best_candidate['FinalAutoMLJobObjectiveMetric']['MetricName'])\n",
    "print(\"FinalAutoMLJobObjectiveMetricValue: \" + str(best_candidate['FinalAutoMLJobObjectiveMetric']['Value']))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model ARN corresponding to the best candidate is : arn:aws:sagemaker:us-east-1:389535300735:model/automl-23-15-11-31\n"
     ]
    }
   ],
   "source": [
    "# Let's now use boto3 to deploy the model\n",
    "model_name = 'automl-' + timestamp_suffix\n",
    "\n",
    "model = sm.create_model(Containers=best_candidate['InferenceContainers'],\n",
    "                            ModelName=model_name,\n",
    "                            ExecutionRoleArn=role)\n",
    "\n",
    "print('Model ARN corresponding to the best candidate is : {}'.format(model['ModelArn']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "automl-endpoint-config2019-12-23-17-15-43\n",
      "Endpoint Config Arn: arn:aws:sagemaker:us-east-1:389535300735:endpoint-config/automl-endpoint-config2019-12-23-17-15-43\n"
     ]
    }
   ],
   "source": [
    "endpoint_config_name = 'automl-endpoint-config' + strftime(\"%Y-%m-%d-%H-%M-%S\", gmtime())\n",
    "print(endpoint_config_name)\n",
    "create_endpoint_config_response = sm.create_endpoint_config(\n",
    "    EndpointConfigName = endpoint_config_name,\n",
    "    ProductionVariants=[{\n",
    "        'InstanceType':'ml.m4.xlarge',\n",
    "        'InitialVariantWeight':1,\n",
    "        'InitialInstanceCount':1,\n",
    "        'ModelName':model_name,\n",
    "        'VariantName':'AllTraffic'}])\n",
    "\n",
    "print(\"Endpoint Config Arn: \" + create_endpoint_config_response['EndpointConfigArn'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create the Endpoint for the AutoML job"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "autmlEndpoint-2019-12-23-17-15-48\n",
      "arn:aws:sagemaker:us-east-1:389535300735:endpoint/autmlendpoint-2019-12-23-17-15-48\n",
      "Status: Creating\n",
      "Status: Creating\n",
      "Status: Creating\n",
      "Status: Creating\n",
      "Status: Creating\n",
      "Status: Creating\n",
      "Status: Creating\n",
      "Status: Creating\n",
      "Status: Creating\n",
      "Status: Creating\n",
      "Status: Creating\n",
      "Status: InService\n",
      "Arn: arn:aws:sagemaker:us-east-1:389535300735:endpoint/autmlendpoint-2019-12-23-17-15-48\n",
      "Status: InService\n",
      "CPU times: user 149 ms, sys: 10.5 ms, total: 159 ms\n",
      "Wall time: 11min 1s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "import time\n",
    "\n",
    "endpoint_name = 'autmlEndpoint-' + strftime(\"%Y-%m-%d-%H-%M-%S\", gmtime())\n",
    "print(endpoint_name)\n",
    "create_endpoint_response = sm.create_endpoint(\n",
    "    EndpointName=endpoint_name,\n",
    "    EndpointConfigName=endpoint_config_name)\n",
    "print(create_endpoint_response['EndpointArn'])\n",
    "\n",
    "resp = sm.describe_endpoint(EndpointName=endpoint_name)\n",
    "status = resp['EndpointStatus']\n",
    "print(\"Status: \" + status)\n",
    "\n",
    "while status=='Creating':\n",
    "    time.sleep(60)\n",
    "    resp = sm.describe_endpoint(EndpointName=endpoint_name)\n",
    "    status = resp['EndpointStatus']\n",
    "    print(\"Status: \" + status)\n",
    "\n",
    "print(\"Arn: \" + resp['EndpointArn'])\n",
    "print(\"Status: \" + status)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test the Model predictions\n",
    "\n",
    "Replace the endpoint name below with your endpoint name from the above cell: it should start with 'autmlEndpoint-yyyy-mm-dd-hh-mm-ss'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sagemaker.predictor import RealTimePredictor, csv_deserializer, csv_serializer\n",
    "endpoint  = 'autmlEndpoint-2019-12-23-17-15-48' #replace with your endpoint name here\n",
    "autopredictor = RealTimePredictor(endpoint = endpoint, serializer=csv_serializer, deserializer =csv_deserializer, \n",
    "                                  content_type='text/csv', sagemaker_session = session)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test = test_sample.Level\n",
    "X_pred = test_sample.drop(columns = ['Level', 'Accessible']).reset_index().drop(columns = ['index'])\n",
    "y_pred = [int(autopredictor.predict(np.array(X_pred.loc[x]))[0][0]) for x in range(len(X_pred))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy from AutoML Job = 0.8148148148148148\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "print(\"Accuracy from AutoML Job = {}\".format(accuracy_score(y_test, y_pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Deployment\n",
    "\n",
    "Here We explore two modes of model deployment -- batch and as a live inference endpoint. \n",
    "First we deploy as a live endpoint and obtain model metrics <br/>\n",
    "Next we do a batch transform job. For this we need to load the test data into S3 as well <br/>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Live Model Deployment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------------------------------------------------------------------------------------------------------------!"
     ]
    }
   ],
   "source": [
    "xgb_predictor = xgb.deploy(initial_instance_count=1,\n",
    "                                 instance_type='ml.m4.xlarge')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb_predictor.content_type = 'text/csv'\n",
    "xgb_predictor.serializer = csv_serializer\n",
    "xgb_predictor.deserializer = json_deserializer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = []\n",
    "y_test = X_test.Level\n",
    "X_test_drop = X_test.drop(columns = 'Level')\n",
    "for i in range(X_test.shape[0]):\n",
    "    predictions.append(xgb_predictor.predict(np.array(X_test_drop.iloc[i])))\n",
    "predictions = np.array(np.round(predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.88      0.51      0.65       274\n",
      "         1.0       0.82      0.97      0.89       639\n",
      "\n",
      "   micro avg       0.83      0.83      0.83       913\n",
      "   macro avg       0.85      0.74      0.77       913\n",
      "weighted avg       0.84      0.83      0.82       913\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, auc, classification_report\n",
    "print(classification_report(y_test, predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8324205914567361"
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_test,predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Notice that the model where we have normalized the features returns a higher accuracy score, but the AutoML does a really good job without virtually any data preparation!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Make Inferences directly in SQL using Amazon Athena"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we will ingest the training data csv file into Amazon Athena and perform inferences directly against the XGBoost SageMaker endpoint. Here we will show how the process works for the XGBoost endpoint. \n",
    "\n",
    "To do so, navigate to SageMaker console and go to Endpoints. Identify the endpoint you just created for XGBoost. It should start with 'sagemaker-xgboost-####'. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Delete Endpoints (Only do so after completing the Athena portion)\n",
    "\n",
    "See the Readme in the Github for this workshop for next steps on how to complete invoke the model endpoints from Amazon Athena"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb_predictor.delete_endpoint()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "autopredictor.delete_endpoint()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "conda_python3",
   "language": "python",
   "name": "conda_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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}