Mar 21 TPC - Raw Feature Importance with Featurewiz.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Introduction\n",
    "\n",
    "This notebook is intended to extract useful insights for the datasets of ‘Tabular Playground Series - Mar 2021’ competition in Kaggle. For this competition, it is required to tackle the Classification problem to predict a continuous target based on a number of feature columns given in the data. All of the feature columns, cat0 - cat9 are categorical, and the feature columns cont0 - cont13 are continuous.\n",
    "\n",
    "We are going to do the feature importance detection for raw features only, using Featurewiz (https://github.com/AutoViML/featurewiz)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!pip install featurewiz"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Initial Preparations\n",
    "\n",
    "We are going to start with the essential pre-requisites as follows\n",
    "\n",
    "- installing *featurewiz* into this notebook\n",
    "- importing the standard Python packages we need to use down the road\n",
    "- programming the useful automation routines for repeatable data visualizations we are going to draw in the Advance Analytical EDA trials down the road"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\numpy\\_distributor_init.py:32: UserWarning: loaded more than 1 DLL from .libs:\n",
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\numpy\\.libs\\libopenblas.IPBC74C7KURV7CB2PKT5Z5FNR3SIBV4J.gfortran-win_amd64.dll\n",
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\numpy\\.libs\\libopenblas.NOIJJG62EMASZI6NYURL6JBKM4EVBGM7.gfortran-win_amd64.dll\n",
      "  stacklevel=1)\n",
      "\n",
      "Bad key \"text.kerning_factor\" on line 4 in\n",
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\matplotlib\\mpl-data\\stylelib\\_classic_test_patch.mplstyle.\n",
      "You probably need to get an updated matplotlibrc file from\n",
      "http://github.com/matplotlib/matplotlib/blob/master/matplotlibrc.template\n",
      "or from the matplotlib source distribution\n"
     ]
    }
   ],
   "source": [
    "import numpy as np # linear algebra\n",
    "import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n",
    "import datetime as dt\n",
    "from typing import Tuple, List, Dict\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import plotly.graph_objects as go\n",
    "from plotly.subplots import make_subplots\n",
    "import plotly.express as px\n",
    "import plotly.offline\n",
    "\n",
    "\n",
    "# read data\n",
    "in_kaggle = False\n",
    "\n",
    "def get_data_file_path(is_in_kaggle: bool) -> Tuple[str, str, str]:\n",
    "    train_path = ''\n",
    "    test_path = ''\n",
    "    sample_submission_path = ''\n",
    "\n",
    "    if is_in_kaggle:\n",
    "        # running in Kaggle, inside the competition\n",
    "        train_path = '../input/tabular-playground-series-mar-2021/train.csv'\n",
    "        test_path = '../input/tabular-playground-series-mar-2021/test.csv'\n",
    "        sample_submission_path = '../input/tabular-playground-series-mar-2021/sample_submission.csv'\n",
    "    else:\n",
    "        # running locally\n",
    "        train_path = 'data/train.csv'\n",
    "        test_path = 'data/test.csv'\n",
    "        sample_submission_path = 'data/sample_submission.csv'\n",
    "\n",
    "    return train_path, test_path, sample_submission_path\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Started at  2021-04-18 23:10:39.923382\n"
     ]
    }
   ],
   "source": [
    "# main flow\n",
    "start_time = dt.datetime.now()\n",
    "print(\"Started at \", start_time)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 1.66 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# get the training set and labels\n",
    "train_set_path, test_set_path, sample_subm_path = get_data_file_path(in_kaggle)\n",
    "\n",
    "df_train = pd.read_csv(train_set_path)\n",
    "df_test = pd.read_csv(test_set_path)\n",
    "\n",
    "subm = pd.read_csv(sample_subm_path)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Basic Data Overview"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 300000 entries, 0 to 299999\n",
      "Data columns (total 32 columns):\n",
      " #   Column  Non-Null Count   Dtype  \n",
      "---  ------  --------------   -----  \n",
      " 0   id      300000 non-null  int64  \n",
      " 1   cat0    300000 non-null  object \n",
      " 2   cat1    300000 non-null  object \n",
      " 3   cat2    300000 non-null  object \n",
      " 4   cat3    300000 non-null  object \n",
      " 5   cat4    300000 non-null  object \n",
      " 6   cat5    300000 non-null  object \n",
      " 7   cat6    300000 non-null  object \n",
      " 8   cat7    300000 non-null  object \n",
      " 9   cat8    300000 non-null  object \n",
      " 10  cat9    300000 non-null  object \n",
      " 11  cat10   300000 non-null  object \n",
      " 12  cat11   300000 non-null  object \n",
      " 13  cat12   300000 non-null  object \n",
      " 14  cat13   300000 non-null  object \n",
      " 15  cat14   300000 non-null  object \n",
      " 16  cat15   300000 non-null  object \n",
      " 17  cat16   300000 non-null  object \n",
      " 18  cat17   300000 non-null  object \n",
      " 19  cat18   300000 non-null  object \n",
      " 20  cont0   300000 non-null  float64\n",
      " 21  cont1   300000 non-null  float64\n",
      " 22  cont2   300000 non-null  float64\n",
      " 23  cont3   300000 non-null  float64\n",
      " 24  cont4   300000 non-null  float64\n",
      " 25  cont5   300000 non-null  float64\n",
      " 26  cont6   300000 non-null  float64\n",
      " 27  cont7   300000 non-null  float64\n",
      " 28  cont8   300000 non-null  float64\n",
      " 29  cont9   300000 non-null  float64\n",
      " 30  cont10  300000 non-null  float64\n",
      " 31  target  300000 non-null  int64  \n",
      "dtypes: float64(11), int64(2), object(19)\n",
      "memory usage: 73.2+ MB\n"
     ]
    }
   ],
   "source": [
    "df_train.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Importance of Raw Features\n",
    "\n",
    "We are going to invoke *featurewiz*, one of the prominent freeware Pythonic AutoML Feature Selection tools, to quickly identify the essential predictors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Imported featurewiz: Auto_ViML's feature engg and selection library. Version=0.0.18\n",
      "output = featurewiz(dataname, target, corr_limit=0.70,\n",
      "                    verbose=2, sep=',', header=0, test_data='',\n",
      "                    feature_engg='', category_encoders='')\n",
      "Let featurewiz add features to your data! Set 'feature_engg' as: 'interactions' or 'groupby' or 'target'\n",
      "                                \n",
      "Skipping feature engineering since no feature_engg input...\n",
      "Shape of your Data Set loaded: (300000, 32)\n",
      "Number of GPUs = 1\n",
      "    No GPU active on this device\n",
      "    Running XGBoost using CPU parameters\n",
      "############## C L A S S I F Y I N G  V A R I A B L E S  ####################\n",
      "Classifying variables in data set...\n",
      "    31 Predictors classified...\n",
      "        5 variable(s) removed since they were ID or low-information variables\n",
      "#### Single_Label Binary_Classification Feature Selection Started ####\n",
      "Starting feature engineering...this will take time...\n",
      "Source X_train shape:  (300000, 26) | X_test shape:  (200000, 26)\n",
      "##################################################\n",
      "> Start preprocessing with 26 variables\n",
      "> Generate Categorical Encoded features\n",
      " + To know more, click: https://github.com/Alex-Lekov/AutoML_Alex/blob/master/automl_alex/encoders.py\n",
      " + added  0  additional Features using FrequencyEncoder\n",
      "> Cleaned NaNs in numeric features\n",
      "##################################################\n",
      "> Final Number of Features:  26\n",
      "##################################################\n",
      "New X_train shape:  (300000, 26) | X_test shape:  (200000, 41)\n",
      "    Completed feature engineering. Shape of Train (with target) = (300000, 27)\n",
      "Searching for highly correlated variables from 26 variables using SULOV method\n",
      "#####  SULOV : Searching for Uncorrelated List Of Variables (takes time...) ############\n",
      "    Removing (3) highly correlated variables:\n",
      "    ['cont0', 'cont2', 'cont7']\n",
      "    Following (23) vars selected: ['cont3', 'cont4', 'cont5', 'cont6', 'cont8', 'cont9', 'FrequencyEncoder_cat17', 'FrequencyEncoder_cat18', 'FrequencyEncoder_cat4', 'FrequencyEncoder_cat16', 'FrequencyEncoder_cat0', 'FrequencyEncoder_cat14', 'FrequencyEncoder_cat12', 'FrequencyEncoder_cat3', 'FrequencyEncoder_cat9', 'FrequencyEncoder_cat13', 'FrequencyEncoder_cat15', 'FrequencyEncoder_cat6', 'FrequencyEncoder_cat11', 'FrequencyEncoder_cat1', 'FrequencyEncoder_cat2', 'cont1', 'cont10']\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    Adding 0 categorical variables to reduced numeric variables  of 23\n",
      "############## F E A T U R E   S E L E C T I O N  ####################\n",
      "Current number of predictors = 23 \n",
      "    Finding Important Features using Boosted Trees algorithm...\n",
      "        using 23 variables...\n",
      "        using 18 variables...\n",
      "        using 13 variables...\n",
      "        using 8 variables...\n",
      "        using 3 variables...\n",
      "Selected 15 important features from your dataset\n",
      "    Time taken (in seconds) = 140\n",
      "Returning list of 15 important features and dataframe.\n",
      "Returning 2 dataframes: train and test with 15 important features.\n"
     ]
    }
   ],
   "source": [
    "from featurewiz import featurewiz\n",
    "\n",
    "# out1 and out2: is always a tuple containing two objects. \n",
    "# In our case, it is going to be \"test\" and \"train\" dataframes with the important features selected by featurewiz\n",
    "trainm, testm = featurewiz(\n",
    "    df_train, \n",
    "    target='target', \n",
    "    corr_limit=0.7, \n",
    "    verbose=2, \n",
    "    sep=\",\",\n",
    "    header=0,\n",
    "    test_data=df_test,     \n",
    "    feature_engg=\"\", # we do not want featurewiz to create any additional features in this experiment\n",
    "    category_encoders=\"FrequencyEncoder\" #OrdinalEncoder\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['FrequencyEncoder_cat16', 'FrequencyEncoder_cat15',\n",
       "       'FrequencyEncoder_cat18', 'FrequencyEncoder_cat0',\n",
       "       'FrequencyEncoder_cat14', 'FrequencyEncoder_cat11',\n",
       "       'FrequencyEncoder_cat13', 'FrequencyEncoder_cat17',\n",
       "       'FrequencyEncoder_cat1', 'cont5', 'FrequencyEncoder_cat4',\n",
       "       'FrequencyEncoder_cat2', 'cont1', 'FrequencyEncoder_cat6', 'cont10',\n",
       "       'target'],\n",
       "      dtype='object')"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# check the important features detected by featurewiz\n",
    "display(trainm.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "We are done. That is all, folks!\n",
      "Finished at  2021-04-18 23:13:02.902602\n",
      "Elapsed time:  0:02:22.979220\n"
     ]
    }
   ],
   "source": [
    "print('We are done. That is all, folks!')\n",
    "finish_time = dt.datetime.now()\n",
    "print(\"Finished at \", finish_time)\n",
    "elapsed = finish_time - start_time\n",
    "print(\"Elapsed time: \", elapsed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Results of Experiments\n",
    "\n",
    "Below are the important raw features selected by featurewiz under different category encoding techniques applied\n",
    "- Cats without encoding: ('cat16', 'cat15', 'cat0', 'cat18', 'cat14', 'cat11', 'cat13', 'cat17', 'cat4', 'cat1', 'cat2', 'cat12')\n",
    "- OrdinalEncoder applied: ('OrdinalEncoder_cat16', 'OrdinalEncoder_cat15', 'OrdinalEncoder_cat0', 'OrdinalEncoder_cat18', 'OrdinalEncoder_cat14', 'OrdinalEncoder_cat11', 'OrdinalEncoder_cat13', 'OrdinalEncoder_cat1', 'OrdinalEncoder_cat17', 'cont5', 'OrdinalEncoder_cat4', 'cont1', 'OrdinalEncoder_cat2', 'cont10')\n",
    "- FrequencyEncoder applied: ('FrequencyEncoder_cat16', 'FrequencyEncoder_cat15', 'FrequencyEncoder_cat18', 'FrequencyEncoder_cat0','FrequencyEncoder_cat14', 'FrequencyEncoder_cat11', 'FrequencyEncoder_cat13', 'FrequencyEncoder_cat17','FrequencyEncoder_cat1', 'cont5', 'FrequencyEncoder_cat4','FrequencyEncoder_cat2', 'cont1', 'FrequencyEncoder_cat6', 'cont10')\n",
    "\n",
    "As we can see, the same set of raw features selected for the final model in the cases of OrdinalEncoder and FrequencyEncoder applied. However, in the case when we relied on the default category handler of featurewiz, it did not detect any continual variable to be important at the end."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# References\n",
    "\n",
    "- [featurewiz repo](https://github.com/AutoViML/featurewiz)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}