1 Cleaning.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Table of Contents"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Context](#Co)<br>\n",
    "        \n",
    "0. [**Merging data files**](#0)<br>\n",
    "    [Import](#0.1)<br>\n",
    "    [Columns comparison](#0.2)<br>\n",
    "    [Rows comparison](#0.3) <br>\n",
    "    [Last adjustments](#0.4) <br>\n",
    "    \n",
    "    \n",
    "1. [**Business validation**](#1)<br>\n",
    "    [Rules and Constraints](#1.1) <br>\n",
    "    [Accuracy](#1.2) <br>\n",
    "    [Consistency](#1.3) <br>\n",
    "    [Uniformity](#1.4) <br>\n",
    "    \n",
    "    \n",
    "2. [**Analytic validation**](#2) <br>\n",
    "    [Irregular Data](#1.4) <br>\n",
    "    [Unnecessary Data](#1.4) <br>    \n",
    "    \n",
    "    \n",
    "3. [**Final report**](#3)\n",
    "    \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"Co\"></a>\n",
    "# Context\n",
    "\n",
    "The **World Happiness Report** is a landmark survey of the state of global happiness from 2015 to 2019 in the world according to 6 main factors:\n",
    "* economic production, \n",
    "* social support, \n",
    "* life expectancy, freedom, \n",
    "* absence of corruption, \n",
    "* and generosity\n",
    "\n",
    "### Purposes of the project\n",
    "<ins> Data analysis: </ins>\n",
    "1. Give a clear picture of happiness around the world in 2019\n",
    "2. Analyse trends in happiness from 2015 to 2019\n",
    "\n",
    "<ins> Forecasting with Machine Learning</ins>(\\*)\n",
    "1. Can we predict a country happiness if we know the gdp per capita, life expectancy and other factors values?\n",
    "2. Can we predict a country happiness thanks to its history (happiness+factors)?\n",
    "\n",
    "(\\*) *Although data don't contain related information, the global pandemic may have a tremendous impact on the results*\n",
    "\n",
    "\n",
    "### Workflow\n",
    "* **Cleaning**\n",
    "* EDA\n",
    "* Data Visualization\n",
    "* Features Engineering\n",
    "* Machine Learning\n",
    "\n",
    "---------\n",
    "\n",
    "Cleaning is the process of detecting and correcting (or removing) corrupt or inaccurate records from the datato ensure that further analysis will be based on meaningful and relevant information.\n",
    "\n",
    "In this notebook, we will go through business and analytical data tests.\n",
    "\n",
    "*(Only tests related to the data will be mentioned but you can find a more exhaustiv list of tests [here](https://docs.google.com/document/d/1YCiHygWCx6A3R4XvKI8LnHtNd4yFUgSbumGx9fVAxXI/edit#heading=h.wvf1xn11p9y0))*\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "------------\n",
    "------------\n",
    "<a name=\"0\"></a>\n",
    "# 0. Merging data files\n",
    "\n",
    "Since our data are split in in 5 files (one for each year of study), our first task is to import the different files and ensure that their information can be matched in order to gather it all in a single dataset."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"0.1\"></a>\n",
    "## Import"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's first import the required packages to import, investigate and plot data: **pandas**, **numpy** and **matplotlib**µ"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from matplotlib import pyplot\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We now need to import the 5 data files. For convenience, we'll store them into a dictionnary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# We create a dictionnary where we'll store our 5 DataFrames, imported from data files\n",
    "dict_data = {}\n",
    "\n",
    "# for each year of study, store data as DataFrame in the dict\n",
    "for year in range(2015, 2020):\n",
    "    df = pd.read_csv('data/'+str(year)+'.csv')\n",
    "    dict_data[year] = df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's have a quick overview of rows and columns for each file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>2015</th>\n",
       "      <th>2016</th>\n",
       "      <th>2017</th>\n",
       "      <th>2018</th>\n",
       "      <th>2019</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>row</th>\n",
       "      <td>158</td>\n",
       "      <td>157</td>\n",
       "      <td>155</td>\n",
       "      <td>156</td>\n",
       "      <td>156</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>col</th>\n",
       "      <td>12</td>\n",
       "      <td>13</td>\n",
       "      <td>12</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     2015  2016  2017  2018  2019\n",
       "row   158   157   155   156   156\n",
       "col    12    13    12     9     9"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# We create a DataFrame where we'll store row and col numbers for each year of study\n",
    "df_dim = pd.DataFrame(['row','col'])\n",
    "# for each year, store row and col numbers in the DataFrame\n",
    "for year in range(2015, 2020):\n",
    "    df_dim = pd.concat([df_dim, pd.DataFrame(dict_data[year].shape, columns=[str(year)])],axis=1)\n",
    "# set index\n",
    "df_dim.set_index(0, inplace=True)\n",
    "df_dim.index.name = None\n",
    "\n",
    "display(df_dim)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see, datasets differs in rows and columns. Thus, we now have to investigate and find a common structure to gather them."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---------------\n",
    "<a name=\"0.2\"></a>\n",
    "## Columns comparison"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, we need to try to identify the features shared/not shared by all the datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>2015</th>\n",
       "      <th>2016</th>\n",
       "      <th>2017</th>\n",
       "      <th>2018</th>\n",
       "      <th>2019</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Country</td>\n",
       "      <td>Country</td>\n",
       "      <td>Country</td>\n",
       "      <td>Overall rank</td>\n",
       "      <td>Overall rank</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Region</td>\n",
       "      <td>Region</td>\n",
       "      <td>Happiness.Rank</td>\n",
       "      <td>Country or region</td>\n",
       "      <td>Country or region</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Happiness Rank</td>\n",
       "      <td>Happiness Rank</td>\n",
       "      <td>Happiness.Score</td>\n",
       "      <td>Score</td>\n",
       "      <td>Score</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Happiness Score</td>\n",
       "      <td>Happiness Score</td>\n",
       "      <td>Whisker.high</td>\n",
       "      <td>GDP per capita</td>\n",
       "      <td>GDP per capita</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Standard Error</td>\n",
       "      <td>Lower Confidence Interval</td>\n",
       "      <td>Whisker.low</td>\n",
       "      <td>Social support</td>\n",
       "      <td>Social support</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Economy (GDP per Capita)</td>\n",
       "      <td>Upper Confidence Interval</td>\n",
       "      <td>Economy..GDP.per.Capita.</td>\n",
       "      <td>Healthy life expectancy</td>\n",
       "      <td>Healthy life expectancy</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Family</td>\n",
       "      <td>Economy (GDP per Capita)</td>\n",
       "      <td>Family</td>\n",
       "      <td>Freedom to make life choices</td>\n",
       "      <td>Freedom to make life choices</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Health (Life Expectancy)</td>\n",
       "      <td>Family</td>\n",
       "      <td>Health..Life.Expectancy.</td>\n",
       "      <td>Generosity</td>\n",
       "      <td>Generosity</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Freedom</td>\n",
       "      <td>Health (Life Expectancy)</td>\n",
       "      <td>Freedom</td>\n",
       "      <td>Perceptions of corruption</td>\n",
       "      <td>Perceptions of corruption</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Trust (Government Corruption)</td>\n",
       "      <td>Freedom</td>\n",
       "      <td>Generosity</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Generosity</td>\n",
       "      <td>Trust (Government Corruption)</td>\n",
       "      <td>Trust..Government.Corruption.</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Dystopia Residual</td>\n",
       "      <td>Generosity</td>\n",
       "      <td>Dystopia.Residual</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>NaN</td>\n",
       "      <td>Dystopia Residual</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                             2015                           2016  \\\n",
       "0                         Country                        Country   \n",
       "1                          Region                         Region   \n",
       "2                  Happiness Rank                 Happiness Rank   \n",
       "3                 Happiness Score                Happiness Score   \n",
       "4                  Standard Error      Lower Confidence Interval   \n",
       "5        Economy (GDP per Capita)      Upper Confidence Interval   \n",
       "6                          Family       Economy (GDP per Capita)   \n",
       "7        Health (Life Expectancy)                         Family   \n",
       "8                         Freedom       Health (Life Expectancy)   \n",
       "9   Trust (Government Corruption)                        Freedom   \n",
       "10                     Generosity  Trust (Government Corruption)   \n",
       "11              Dystopia Residual                     Generosity   \n",
       "12                            NaN              Dystopia Residual   \n",
       "\n",
       "                             2017                          2018  \\\n",
       "0                         Country                  Overall rank   \n",
       "1                  Happiness.Rank             Country or region   \n",
       "2                 Happiness.Score                         Score   \n",
       "3                    Whisker.high                GDP per capita   \n",
       "4                     Whisker.low                Social support   \n",
       "5        Economy..GDP.per.Capita.       Healthy life expectancy   \n",
       "6                          Family  Freedom to make life choices   \n",
       "7        Health..Life.Expectancy.                    Generosity   \n",
       "8                         Freedom     Perceptions of corruption   \n",
       "9                      Generosity                           NaN   \n",
       "10  Trust..Government.Corruption.                           NaN   \n",
       "11              Dystopia.Residual                           NaN   \n",
       "12                            NaN                           NaN   \n",
       "\n",
       "                            2019  \n",
       "0                   Overall rank  \n",
       "1              Country or region  \n",
       "2                          Score  \n",
       "3                 GDP per capita  \n",
       "4                 Social support  \n",
       "5        Healthy life expectancy  \n",
       "6   Freedom to make life choices  \n",
       "7                     Generosity  \n",
       "8      Perceptions of corruption  \n",
       "9                            NaN  \n",
       "10                           NaN  \n",
       "11                           NaN  \n",
       "12                           NaN  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# We create a DataFrame to store the columns lists for each set.\n",
    "df_col = pd.DataFrame()\n",
    "# For each year, store columns in DataFrame\n",
    "for year in range(2015, 2020):\n",
    "    df_col = pd.concat([df_col, pd.DataFrame(dict_data[year].columns.tolist(), columns=[str(year)])],axis=1)\n",
    "\n",
    "display(df_col)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "OK. Looking at this output, we can easily notice 3 different cases:\n",
    "\n",
    "**1. Features shared by the 5 sets with the  same (or slightly different) names**\n",
    "- Happiness Rank / Happiness.Rank / Overall rank\n",
    "- Happiness Score/ Happiness.Score / Happiness.Score / Score\n",
    "- Economy (GDP per Capita) / Economy..GDP.per.Capita./ GDP per capita\n",
    "- Health (Life Expectancy) / Health..Life.Expectancy. / Healthy life expectancy\n",
    "- Freedom / Freedom to make life choices\n",
    "- Generosity\n",
    "\n",
    "\n",
    "**2. Features who seem to be the same but need a deeper examination**\n",
    "- Country + Region / Country / Country or region\n",
    "- Trust (Government Corruption)\t/ Trust..Government.Corruption.\t/ Perceptions of corruption\t\n",
    "- Family / Social Support\n",
    "\n",
    "\n",
    "**3. Other features that are not contained in all the data files**\n",
    "- 2015, 2016, 2017: Dystopia Residual / Dystopia.Residual\t\n",
    "- 2015 : Standard error\n",
    "- 2016 : Lower Confidence Interval + Upper Confidence Interval\t\n",
    "- 2017 : Whisker.high + Whisker.low"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----------\n",
    "**1. Features shared by the 5 sets with the  same (or slightly different) names**\n",
    "\n",
    "This case is easy to deal with. We just have to rename the features (in respect of naming convensions: lower case and underscores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# 2015 \n",
    "dict_data[2015].rename(columns={'Happiness Rank': \"happiness_rank\",  \"Happiness Score\": \"happiness_score\", \"Economy (GDP per Capita)\":\"gdp_per_capita\", \n",
    "                                \"Health (Life Expectancy)\": \"life_expectancy\", \"Freedom\": \"freedom\", \"Generosity\": \"generosity\"}, inplace=True)\n",
    "# 2016\n",
    "dict_data[2016].rename(columns={'Happiness Rank': \"happiness_rank\", \"Happiness Score\": \"happiness_score\", \"Economy (GDP per Capita)\":\"gdp_per_capita\", \n",
    "                                \"Health (Life Expectancy)\": \"life_expectancy\", \"Freedom\": \"freedom\", \"Generosity\": \"generosity\"}, inplace=True)\n",
    "# 2017\n",
    "dict_data[2017].rename(columns={'Happiness.Rank': \"happiness_rank\", \"Happiness.Score\": \"happiness_score\", \"Economy..GDP.per.Capita.\":\"gdp_per_capita\", \n",
    "                                \"Health..Life.Expectancy.\": \"life_expectancy\", \"Freedom\": \"freedom\", \"Generosity\": \"generosity\"}, inplace=True)\n",
    "# 2018\n",
    "dict_data[2018].rename(columns={'Overall rank': \"happiness_rank\", \"Score\": \"happiness_score\", \"GDP per capita\":\"gdp_per_capita\", \n",
    "                                \"Healthy life expectancy\": \"life_expectancy\", \"Freedom to make life choices\": \"freedom\", \"Generosity\": \"generosity\"}, inplace=True)\n",
    "# 2019\n",
    "dict_data[2019].rename(columns={'Overall rank': \"happiness_rank\", \"Score\": \"happiness_score\", \"GDP per capita\":\"gdp_per_capita\",  \n",
    "                                \"Healthy life expectancy\": \"life_expectancy\", \"Freedom to make life choices\": \"freedom\", \"Generosity\": \"generosity\"}, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**2. Features who seem to be the same but need a deeper examination**\n",
    "\n",
    "For the following features, we need to go deeper to make sure they contain the same information"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Country + Region / Country / Country or Region "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141 are present in the 5 datasets\n"
     ]
    }
   ],
   "source": [
    "# First, let's store countries values in dict\n",
    "d_countries = {}\n",
    "d_countries[2015] = dict_data[2015]['Country'].tolist()\n",
    "d_countries[2016] = dict_data[2016]['Country'].tolist()\n",
    "d_countries[2017] = dict_data[2017]['Country'].tolist()\n",
    "d_countries[2018] = dict_data[2018]['Country or region'].tolist()\n",
    "d_countries[2019] = dict_data[2019]['Country or region'].tolist()\n",
    "\n",
    "# Identify the countries that appear in all sets.\n",
    "l_common_countries = set(d_countries[2015]).intersection(d_countries[2016], d_countries[2017], d_countries[2018], d_countries[2019])\n",
    "print(str(len(l_common_countries))+\" are present in the 5 datasets\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since we know that sets contains between 155 and 158 rows each, we can assert that all the features contains the same information"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "------------\n",
    "* Trust (Government Corruption) / Trust..Government.Corruption. / Perceptions of corruption\n",
    "\n",
    "Let's compare the distribution in the 5 sets:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x216 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# First, let's store government corruption values in dict\n",
    "d_gov = {}\n",
    "d_gov[2015] = dict_data[2015]['Trust (Government Corruption)'].tolist()\n",
    "d_gov[2016] = dict_data[2016]['Trust (Government Corruption)'].tolist()\n",
    "d_gov[2017] = dict_data[2017]['Trust..Government.Corruption.'].tolist()\n",
    "d_gov[2018] = dict_data[2018]['Perceptions of corruption'].fillna(dict_data[2018]['Perceptions of corruption'].median())\n",
    "d_gov[2019] = dict_data[2019]['Perceptions of corruption']\n",
    "\n",
    "# Boxplot\n",
    "pyplot.figure(figsize=(15,3))\n",
    "pyplot.boxplot([d_gov[2015], d_gov[2016], d_gov[2017], d_gov[2018], d_gov[2019]])\n",
    "pyplot.xticks([1, 2, 3, 4, 5], ['2015', '2016', '2017','2018','2019'])\n",
    "pyplot.title('Trust (Government Corruption) / Trust..Government.Corruption. / Perceptions of corruption')\n",
    "pyplot.show()\n",
    "\n",
    "# Histogram\n",
    "pyplot.figure(figsize=(15,3))\n",
    "i = 151 # subplot initial position\n",
    "for year in range(2015, 2020):\n",
    "    pyplot.subplot(i)\n",
    "    pyplot.hist(d_gov[year])\n",
    "    pyplot.title(str(year))\n",
    "    i+=1\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can observe that distribution over the years is very similar. Again, we can assert the 5 features contain the same information."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----------\n",
    "* Family / Social Support"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x216 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# First, let's store government corruption values in dict\n",
    "d_gov = {}\n",
    "d_gov[2015] = dict_data[2015]['Family'].tolist()\n",
    "d_gov[2016] = dict_data[2016]['Family'].tolist()\n",
    "d_gov[2017] = dict_data[2017]['Family'].tolist()\n",
    "d_gov[2018] = dict_data[2018]['Social support']\n",
    "d_gov[2019] = dict_data[2019]['Social support']\n",
    "\n",
    "# Boxplot\n",
    "pyplot.figure(figsize=(15,3))\n",
    "pyplot.boxplot([d_gov[2015], d_gov[2016], d_gov[2017], d_gov[2018], d_gov[2019]])\n",
    "pyplot.xticks([1, 2, 3, 4, 5], ['2015', '2016', '2017','2018','2019'])\n",
    "pyplot.title('Family / Social Support')\n",
    "pyplot.show()\n",
    "\n",
    "# Histogram\n",
    "pyplot.figure(figsize=(15,3))\n",
    "i = 151 # subplot initial position\n",
    "for year in range(2015, 2020):\n",
    "    pyplot.subplot(i)\n",
    "    pyplot.hist(d_gov[year])\n",
    "    pyplot.title(str(year))\n",
    "    i+=1\n",
    "pyplot.show()\n",
    "\n",
    "# => We can observe that the distribution is very similar between 2017 and (2018-2019). \n",
    "# Thus, we can assume that Social support and family \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looking at the distribution, which is very similar between 2017 and 2018, we can assert that Family and Social support contain the same information.\n",
    " \n",
    "Let's rename the 3 features with common names: `country`, `coruption_perception` and `social_support`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# country\n",
    "dict_data[2015].rename(columns={'Country': \"country\"}, inplace=True)\n",
    "dict_data[2016].rename(columns={'Country': \"country\"}, inplace=True)\n",
    "dict_data[2017].rename(columns={'Country': \"country\"}, inplace=True)\n",
    "dict_data[2018].rename(columns={'Country or region': \"country\"}, inplace=True)\n",
    "dict_data[2019].rename(columns={'Country or region': \"country\"}, inplace=True)\n",
    "\n",
    "# corruption_perception\n",
    "dict_data[2015].rename(columns={'Trust (Government Corruption)': \"corruption_perception\"}, inplace=True)\n",
    "dict_data[2016].rename(columns={'Trust (Government Corruption)': \"corruption_perception\"}, inplace=True)\n",
    "dict_data[2017].rename(columns={'Trust..Government.Corruption.': \"corruption_perception\"}, inplace=True)\n",
    "dict_data[2018].rename(columns={'Perceptions of corruption': \"corruption_perception\"}, inplace=True)\n",
    "dict_data[2019].rename(columns={'Perceptions of corruption': \"corruption_perception\"}, inplace=True)\n",
    "\n",
    "# social_support\n",
    "dict_data[2015].rename(columns={'Family': \"social_support\"}, inplace=True)\n",
    "dict_data[2016].rename(columns={'Family': \"social_support\"}, inplace=True)\n",
    "dict_data[2017].rename(columns={'Family': \"social_support\"}, inplace=True)\n",
    "dict_data[2018].rename(columns={'Social support': \"social_support\"}, inplace=True)\n",
    "dict_data[2019].rename(columns={'Social support': \"social_support\"}, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---------\n",
    "**3. Other features that are not contained in all the data files**\n",
    "\n",
    "Since the purpose of this project is to analyse `happiness`` through time, we take the side of removing these features."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 2015, 2016, 2017: Dystopia Residual / Dystopia.Residual\n",
    "# 2015 : Standard error\n",
    "# 2016 : Lower Confidence Interval + Upper Confidence Interval\n",
    "# 2017 : Whisker.high + Whisker.low\n",
    "dict_data[2015] = dict_data[2015].drop(columns=['Dystopia Residual','Standard Error'])\n",
    "dict_data[2016] = dict_data[2016].drop(columns=['Dystopia Residual','Lower Confidence Interval','Upper Confidence Interval'])\n",
    "dict_data[2017] = dict_data[2017].drop(columns=['Dystopia.Residual','Whisker.high','Whisker.low'])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----------\n",
    "<a name=\"0.3\"></a>\n",
    "## Rows comparison\n",
    "As we saw earleier, 141 countries are common between the 5 sets, let's keep only these ones."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# For each set, we only keep the 141 common countries previously identified\n",
    "for year in range(2015, 2020):\n",
    "    df = dict_data[year]\n",
    "    df = df[df['country'].isin(l_common_countries)]\n",
    "    dict_data[year] = df\n",
    "    \n",
    "# We could look closer to the name and try to find same countries with different orthographes but we won't do that here."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our 5 files now contain the same countries. But because we removed some observations in each, we have to recalculate the `happiness rank`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Done !\n"
     ]
    }
   ],
   "source": [
    "for year in range(2015, 2020):\n",
    "    df = dict_data[year].sort_values(by='happiness_rank')\n",
    "    df['happiness_rank'] = np.arange(len(df))+1\n",
    "    dict_data[year] = df\n",
    "    \n",
    "print(\"Done !\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-------------\n",
    "<a name=\"0.4\"></a>\n",
    "## Last adjustments\n",
    "\n",
    "* Since 2015 and 2016 sets contains corresponding `regions` to `countries`, we can add this information to the others sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(141, 10)\n",
      "(141, 10)\n",
      "(141, 10)\n"
     ]
    }
   ],
   "source": [
    "# Rename Region in respect of naming conventions\n",
    "dict_data[2015].rename(columns={'Region':'region'}, inplace=True)\n",
    "dict_data[2016].rename(columns={'Region':'region'}, inplace=True)\n",
    "\n",
    "# add region information in 2017-2019 sets\n",
    "df_country_reg = dict_data[2015][['country','region']].copy()\n",
    "for year in range(2017,2020):\n",
    "    dict_data[year] = pd.merge(dict_data[year], df_country_reg, on='country', how='inner')\n",
    "    print(dict_data[year].shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* We add a new column containing the `year` of study"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# for each set, create a new column \"year\" containing the year of study\n",
    "for year in range(2015,2020):\n",
    "    if 'year' not in dict_data[year].columns.tolist():\n",
    "        dict_data[year]['year']=year"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "------------\n",
    "## Validation and merge"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "let's take a look at our modified data files to ensure they can be merged"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>2015</th>\n",
       "      <th>2016</th>\n",
       "      <th>2017</th>\n",
       "      <th>2018</th>\n",
       "      <th>2019</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>corruption_perception</td>\n",
       "      <td>corruption_perception</td>\n",
       "      <td>corruption_perception</td>\n",
       "      <td>corruption_perception</td>\n",
       "      <td>corruption_perception</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>country</td>\n",
       "      <td>country</td>\n",
       "      <td>country</td>\n",
       "      <td>country</td>\n",
       "      <td>country</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>freedom</td>\n",
       "      <td>freedom</td>\n",
       "      <td>freedom</td>\n",
       "      <td>freedom</td>\n",
       "      <td>freedom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>gdp_per_capita</td>\n",
       "      <td>gdp_per_capita</td>\n",
       "      <td>gdp_per_capita</td>\n",
       "      <td>gdp_per_capita</td>\n",
       "      <td>gdp_per_capita</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>generosity</td>\n",
       "      <td>generosity</td>\n",
       "      <td>generosity</td>\n",
       "      <td>generosity</td>\n",
       "      <td>generosity</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>happiness_rank</td>\n",
       "      <td>happiness_rank</td>\n",
       "      <td>happiness_rank</td>\n",
       "      <td>happiness_rank</td>\n",
       "      <td>happiness_rank</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>happiness_score</td>\n",
       "      <td>happiness_score</td>\n",
       "      <td>happiness_score</td>\n",
       "      <td>happiness_score</td>\n",
       "      <td>happiness_score</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>life_expectancy</td>\n",
       "      <td>life_expectancy</td>\n",
       "      <td>life_expectancy</td>\n",
       "      <td>life_expectancy</td>\n",
       "      <td>life_expectancy</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>region</td>\n",
       "      <td>region</td>\n",
       "      <td>region</td>\n",
       "      <td>region</td>\n",
       "      <td>region</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>social_support</td>\n",
       "      <td>social_support</td>\n",
       "      <td>social_support</td>\n",
       "      <td>social_support</td>\n",
       "      <td>social_support</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>year</td>\n",
       "      <td>year</td>\n",
       "      <td>year</td>\n",
       "      <td>year</td>\n",
       "      <td>year</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     2015                   2016                   2017  \\\n",
       "0   corruption_perception  corruption_perception  corruption_perception   \n",
       "1                 country                country                country   \n",
       "2                 freedom                freedom                freedom   \n",
       "3          gdp_per_capita         gdp_per_capita         gdp_per_capita   \n",
       "4              generosity             generosity             generosity   \n",
       "5          happiness_rank         happiness_rank         happiness_rank   \n",
       "6         happiness_score        happiness_score        happiness_score   \n",
       "7         life_expectancy        life_expectancy        life_expectancy   \n",
       "8                  region                 region                 region   \n",
       "9          social_support         social_support         social_support   \n",
       "10                   year                   year                   year   \n",
       "\n",
       "                     2018                   2019  \n",
       "0   corruption_perception  corruption_perception  \n",
       "1                 country                country  \n",
       "2                 freedom                freedom  \n",
       "3          gdp_per_capita         gdp_per_capita  \n",
       "4              generosity             generosity  \n",
       "5          happiness_rank         happiness_rank  \n",
       "6         happiness_score        happiness_score  \n",
       "7         life_expectancy        life_expectancy  \n",
       "8                  region                 region  \n",
       "9          social_support         social_support  \n",
       "10                   year                   year  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Same as we did earlier\n",
    "df_col = pd.DataFrame()\n",
    "\n",
    "for year in range(2015, 2020):\n",
    "    dict_data[year] = dict_data[year].reindex(sorted(dict_data[year].columns), axis=1)\n",
    "    df_col = pd.concat([df_col, pd.DataFrame(dict_data[year].columns.tolist(), columns=[str(year)])],axis=1)\n",
    "\n",
    "display(df_col)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The first step of cleaning our data is over. As we can see, our 5 datasets columns match perfectly, we can now merge the 5 sets in a single DataFrame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gathering dataset dimensions: (705, 11)\n"
     ]
    },
    {
     "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>columns</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>corruption_perception</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>country</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>freedom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>gdp_per_capita</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>generosity</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>happiness_rank</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>happiness_score</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>life_expectancy</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>region</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>social_support</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>year</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  columns\n",
       "0   corruption_perception\n",
       "1                 country\n",
       "2                 freedom\n",
       "3          gdp_per_capita\n",
       "4              generosity\n",
       "5          happiness_rank\n",
       "6         happiness_score\n",
       "7         life_expectancy\n",
       "8                  region\n",
       "9          social_support\n",
       "10                   year"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# We create our gathering dataset\n",
    "df = pd.DataFrame()\n",
    "\n",
    "# we store each set in it\n",
    "for year in range(2015,2020):\n",
    "    df = pd.concat([df, dict_data[year]])\n",
    "\n",
    "print(\"gathering dataset dimensions:\",df.shape)\n",
    "display(pd.DataFrame({\"columns\":df.columns.tolist()}))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---------------\n",
    "---------------\n",
    "<a name=\"1\"></a>\n",
    "# 1. Business validation\n",
    "\n",
    "There are a number of ways to examine data sets for errors and anomalies, but all depend upon preconstructed validation rules. These rules set expected parameters for data, and they vary widely in complexity based upon the needs of the user. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"1.1\"></a>\n",
    "## Rules and Constraints\n",
    "Test if the data data conform to defined business rules or constraints.\n",
    "\n",
    "*Since we have no business expert for this dataset, we'll try to figure out our propre rules.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Data types**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "corruption_perception    float64\n",
       "country                   object\n",
       "freedom                  float64\n",
       "gdp_per_capita           float64\n",
       "generosity               float64\n",
       "happiness_rank             int32\n",
       "happiness_score          float64\n",
       "life_expectancy          float64\n",
       "region                    object\n",
       "social_support           float64\n",
       "year                       int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can expect:\n",
    "* `country` and `region are categorical features\n",
    "* `year` and `happiness rank` are integer numbers\n",
    "* `happiness score` and `factors` are float numbers (decimal)\n",
    "\n",
    "Nothing to change here."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Range**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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>corruption_perception</th>\n",
       "      <th>freedom</th>\n",
       "      <th>gdp_per_capita</th>\n",
       "      <th>generosity</th>\n",
       "      <th>happiness_rank</th>\n",
       "      <th>happiness_score</th>\n",
       "      <th>life_expectancy</th>\n",
       "      <th>social_support</th>\n",
       "      <th>year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>2.839</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>2015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.55191</td>\n",
       "      <td>0.724</td>\n",
       "      <td>2.096</td>\n",
       "      <td>0.838075</td>\n",
       "      <td>141</td>\n",
       "      <td>7.769</td>\n",
       "      <td>1.141</td>\n",
       "      <td>1.644</td>\n",
       "      <td>2019</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     corruption_perception  freedom  gdp_per_capita  generosity  \\\n",
       "min                0.00000    0.000           0.000    0.000000   \n",
       "max                0.55191    0.724           2.096    0.838075   \n",
       "\n",
       "     happiness_rank  happiness_score  life_expectancy  social_support  year  \n",
       "min               1            2.839            0.000           0.000  2015  \n",
       "max             141            7.769            1.141           1.644  2019  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# We define a function to compute the min and max of a column\n",
    "def minMax(x):\n",
    "    return pd.Series(index=['min','max'],data=[x.min(),x.max()])\n",
    "\n",
    "# We apply our function to numerica data\n",
    "df._get_numeric_data().apply(minMax)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can expect:\n",
    "* `year` range is between 2015 and 2019\n",
    "* `happiness rank` values are between 1 and 141\n",
    "* `happiness score` and `factors` are >0\n",
    "\n",
    "Nothing to change here."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Mandatory**\n",
    "\n",
    "Some features are required not to be empty. In this dataset, we expect `country` (identifier) and `happinesss core` (project focus) to be filled."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "country            0\n",
       "happiness_score    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[['country','happiness_score']].isna().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`Country` and `happiness score` columns don't contain missing data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "--------------\n",
    "<a name=\"1.2\"></a>\n",
    "## Accuracy\n",
    "We want to make sure the values in the data are true.\n",
    "Since we have no expertise on this dataset, all we can do is to check the `happiness rank` according to the `happiness score`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x216 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Scatter plot\n",
    "pyplot.figure(figsize=(15,3))\n",
    "i = 151 # subplot initial position\n",
    "for year in range(2015, 2020):\n",
    "    pyplot.subplot(i)\n",
    "    df_tmp = df[df['year']==year]\n",
    "    pyplot.scatter(x=df_tmp['happiness_rank'], y=df_tmp['happiness_score'])\n",
    "    pyplot.title(str(year))\n",
    "    i+=1\n",
    "pyplot.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As expected, the higher the score, the better the rank"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"2\"></a>\n",
    "# Analytic validation\n",
    "\n",
    "As well as allowing us to work on more meaningful information, cleaning the data in an analytical way  is mandatory for an efficient use of Data Science algorithms. Thus, it is important to understand how to detect and clean anomalies in order to improve data quality.\n",
    "Inspecting the data is time-consuming and requires using many methods for exploring the underlying data for error detection, such as descriptives statistiques and data visualization.\n",
    "\n",
    "*(Only tests related to our data will be mentioned but you can find the whole list [here](https://docs.google.com/document/d/1YCiHygWCx6A3R4XvKI8LnHtNd4yFUgSbumGx9fVAxXI/edit#heading=h.wvf1xn11p9y0))*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"2.1\"></a>\n",
    "## Irregular data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Missing data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>missing_values</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>corruption_perception</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       missing_values\n",
       "corruption_perception               1"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>corruption_perception</th>\n",
       "      <th>country</th>\n",
       "      <th>freedom</th>\n",
       "      <th>gdp_per_capita</th>\n",
       "      <th>generosity</th>\n",
       "      <th>happiness_rank</th>\n",
       "      <th>happiness_score</th>\n",
       "      <th>life_expectancy</th>\n",
       "      <th>region</th>\n",
       "      <th>social_support</th>\n",
       "      <th>year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>NaN</td>\n",
       "      <td>United Arab Emirates</td>\n",
       "      <td>0.284</td>\n",
       "      <td>2.096</td>\n",
       "      <td>0.186</td>\n",
       "      <td>20</td>\n",
       "      <td>6.774</td>\n",
       "      <td>0.67</td>\n",
       "      <td>Middle East and Northern Africa</td>\n",
       "      <td>0.776</td>\n",
       "      <td>2018</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    corruption_perception               country  freedom  gdp_per_capita  \\\n",
       "19                    NaN  United Arab Emirates    0.284           2.096   \n",
       "\n",
       "    generosity  happiness_rank  happiness_score  life_expectancy  \\\n",
       "19       0.186              20            6.774             0.67   \n",
       "\n",
       "                             region  social_support  year  \n",
       "19  Middle East and Northern Africa           0.776  2018  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Let's plot the features that contains at least 1 missing values\n",
    "na_df = pd.DataFrame({'missing_values':df.isna().sum()})\n",
    "\n",
    "display(na_df[na_df['missing_values']>0])\n",
    "\n",
    "# let's plot the lines concerned by the missing value\n",
    "display(df[df['corruption_perception'].isna()])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see, there is only one missing value: `corruption perception` in 2018 for United Arab Emirates. The best way to input it is to use the value from 2017."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Done!\n"
     ]
    }
   ],
   "source": [
    "# get corruption_perception values for the year 2017\n",
    "df_corrupt_2017 = dict_data[2017][['country', 'corruption_perception']].rename(columns={'corruption_perception':'corr_2017'})\n",
    "\n",
    "# merge\n",
    "df_temp =pd.merge(df, df_corrupt_2017, on='country')\n",
    "\n",
    "# replace value\n",
    "df_temp['corruption_perception'] = df_temp.apply(\n",
    "    lambda row: row['corr_2017'] if np.isnan(row['corruption_perception']) else row['corruption_perception'],\n",
    "    axis=1\n",
    ")\n",
    "\n",
    "df = df_temp.drop(['corr_2017'], axis=1)\n",
    "\n",
    "print(\"Done!\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* outliers:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 7 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 7 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#boxplot\n",
    "pyplot.figure(figsize=(20,3))\n",
    "\n",
    "i=171\n",
    "for col in['corruption_perception','freedom', 'gdp_per_capita', 'generosity', 'happiness_rank', 'life_expectancy', 'social_support']:\n",
    "    pyplot.subplot(i)\n",
    "    pyplot.boxplot(df[col])\n",
    "    pyplot.title(col)\n",
    "    i+=1\n",
    "pyplot.show()\n",
    "\n",
    "pyplot.figure(figsize=(20,3))\n",
    "j = 171\n",
    "for col in['corruption_perception','freedom', 'gdp_per_capita', 'generosity', 'happiness_rank', 'life_expectancy', 'social_support']:\n",
    "    pyplot.subplot(j)\n",
    "    pyplot.hist(df[col])\n",
    "    pyplot.title(col)\n",
    "    j+=1\n",
    "pyplot.show()\n",
    "\n",
    "#pyplot.figure(figsize=(20,3))\n",
    "pd.DataFrame(df['region'].value_counts()).reset_index().rename(columns={'index':'region', 'region':'values_count'}).plot.barh(x='region',y='values_count')\n",
    "pyplot.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can observe that `corruption perception` and `genoristy` features may have some extrem values. We'll use the IQR method to replace them."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of corruption_perception outliers: 67\n",
      "number of generosity outliers: 15\n",
      "\n",
      "==> outliers values have been replace thanks to IQR method\n"
     ]
    }
   ],
   "source": [
    "# We define a function that returns outliers thresholds using IQR\n",
    "def outliers_th(col):\n",
    "    sorted(col)\n",
    "    Q1,Q3 = col.quantile([0.25,0.75])\n",
    "    IQR = Q3 - Q1\n",
    "    lower_range = Q1 - (1.5*IQR)\n",
    "    upper_range = Q3 + (1.5*IQR)\n",
    "    \n",
    "    return lower_range, upper_range\n",
    "\n",
    "lower_corrupt, upper_corrupt = outliers_th(df['corruption_perception'])\n",
    "lower_gen, upper_gen = outliers_th(df['generosity'])\n",
    "\n",
    "\n",
    "# number of outliers for corruption_perception\n",
    "print(\"number of corruption_perception outliers: \"+str((df['corruption_perception']>upper_corrupt).sum()))\n",
    "\n",
    "# number of outliers for generosity\n",
    "print(\"number of generosity outliers: \"+str((df['generosity']>upper_gen).sum()))\n",
    "\n",
    "# replace values\n",
    "df['corruption_perception'] = df.apply(\n",
    "    lambda row: upper_corrupt if row['corruption_perception']>upper_corrupt else row['corruption_perception'],\n",
    "    axis=1)\n",
    "\n",
    "df['generosity'] = df.apply(\n",
    "    lambda row: upper_gen if row['generosity']>upper_gen else row['generosity'],\n",
    "    axis=1)\n",
    "\n",
    "print(\"\\n==> outliers values have been replace thanks to IQR method\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 7 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#boxplot\n",
    "pyplot.figure(figsize=(20,3))\n",
    "\n",
    "i=171\n",
    "for col in['corruption_perception','freedom', 'gdp_per_capita', 'generosity', 'happiness_rank', 'life_expectancy', 'social_support']:\n",
    "    pyplot.subplot(i)\n",
    "    pyplot.boxplot(df[col])\n",
    "    pyplot.title(col)\n",
    "    i+=1\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----------\n",
    "<a name=\"2.2\"></a>\n",
    "## Unnecessary Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Uninformative / repetitive: features with many rows having the same values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 7 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "`\n",
      " Regions:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Sub-Saharan Africa                 155\n",
       "Central and Eastern Europe         140\n",
       "Latin America and Caribbean        100\n",
       "Western Europe                     100\n",
       "Middle East and Northern Africa     95\n",
       "Southeastern Asia                   40\n",
       "Southern Asia                       35\n",
       "Eastern Asia                        20\n",
       "North America                       10\n",
       "Australia and New Zealand           10\n",
       "Name: region, dtype: int64"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pyplot.figure(figsize=(20,3))\n",
    "j = 191\n",
    "for col in [col for col in df.columns.tolist() if col not in ['country','happiness_rank','year','region']]:\n",
    "    pyplot.subplot(j)\n",
    "    pyplot.hist(df[col])\n",
    "    pyplot.title(col)\n",
    "    j+=1\n",
    "pyplot.show()\n",
    "\n",
    "print(\"`\\n Regions:\")\n",
    "display(df[\"region\"].value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see that all of the features are well distributed"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----------\n",
    "<a name=\"2.2\"></a>\n",
    "## Inconsistent Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Data Normalization\n",
    "\n",
    "In order to have a better understanding of the further analysis, and because many machine learning models work better with similar scaled variables, we rescale the different scores (`happiness score` and `factors``) to range 0-100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Rescale\n",
    "df_norm = df.copy()\n",
    "for col in ['corruption_perception', 'freedom', 'gdp_per_capita', \n",
    "            'generosity', 'happiness_score', 'life_expectancy', 'social_support']:\n",
    "        \n",
    "        df_norm[col] = (df_norm[col]-df_norm[col].min())/(df_norm[col].max()-df_norm[col].min())*100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "------------\n",
    "------------\n",
    "<a name=\"3\"></a>\n",
    "# 3. Final report"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The changes we made\n",
    "\n",
    "In this notebook, we cleaned our raw data files to ease and improve the quality of our further analysis.\n",
    "\n",
    "**Our first task was to gather the 5 data files (one for each year of study) into one single dataset. To do so, we had to make some modifications to**\n",
    "* We renamed features that contained same informations but had different names.\n",
    "* We removed features which were not present for all the years of study: Dystopia residual, standard error and confidence interval \n",
    "* We made sure to have the same countries for all the years of study because we aim to look at happiness evolution through time. So we removed about 10% of the countries in each file and kept 141 countries for our final dataset. (This step can be improved by finding countries with different orthographes in the files)\n",
    "* We finally merged the files into a single dataset that we will use for deeper analysis\n",
    "\n",
    "=> The dataset set now contains the following features: `country`, `happiness_score`, `happiness_rank`, `gdp_per_capita`, `life_expectancy`, `social_support`, `freedom`, `corruption_perception`, `generosity`, `region`, `year`\n",
    "\n",
    "**After this step, we corrected the data according to business Expertise. Since we have no real expertise, we tried to figure out our own rules. We looked for different kinds of invalid data:**\n",
    "* Data types\n",
    "* Data ranges\n",
    "* Mandatody values\n",
    "* Accuracy of the value\n",
    "\n",
    "Fortunately, the data respected all our rules and we havn't had to make modifications.\n",
    "\n",
    "**Finally, our last task was to verify the data in an analytical way by looking for anomalies in our features:**\n",
    "* Missing values\n",
    "* Outliers\n",
    "* repetitive data\n",
    "* Data normalization\n",
    "\n",
    "Only one single missing value has been found, and replaced with the mean of the other values for the same feature.\n",
    "\n",
    "## The final dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Our final datasets contains 705 observations and 11 columns\n"
     ]
    },
    {
     "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>Features</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>corruption_perception</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>country</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>freedom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>gdp_per_capita</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>generosity</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>happiness_rank</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>happiness_score</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>life_expectancy</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>region</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>social_support</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>year</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 Features\n",
       "0   corruption_perception\n",
       "1                 country\n",
       "2                 freedom\n",
       "3          gdp_per_capita\n",
       "4              generosity\n",
       "5          happiness_rank\n",
       "6         happiness_score\n",
       "7         life_expectancy\n",
       "8                  region\n",
       "9          social_support\n",
       "10                   year"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(\"Our final datasets contains \"+str(df.shape[0])+' observations and '+str(df.shape[1])+' columns')\n",
    "display(pd.DataFrame({\"Features\":df.columns.tolist()}))\n",
    "\n",
    "# export\n",
    "df.to_csv('data/data_clean.csv', index=False)\n",
    "df_norm.to_csv('data/data_clean_norm.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(705, 11)\n"
     ]
    },
    {
     "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>corruption_perception</th>\n",
       "      <th>country</th>\n",
       "      <th>freedom</th>\n",
       "      <th>gdp_per_capita</th>\n",
       "      <th>generosity</th>\n",
       "      <th>happiness_rank</th>\n",
       "      <th>happiness_score</th>\n",
       "      <th>life_expectancy</th>\n",
       "      <th>region</th>\n",
       "      <th>social_support</th>\n",
       "      <th>year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>262</th>\n",
       "      <td>0.130688</td>\n",
       "      <td>Nicaragua</td>\n",
       "      <td>0.447552</td>\n",
       "      <td>0.737299</td>\n",
       "      <td>0.301674</td>\n",
       "      <td>41</td>\n",
       "      <td>6.071</td>\n",
       "      <td>0.653096</td>\n",
       "      <td>Latin America and Caribbean</td>\n",
       "      <td>1.287216</td>\n",
       "      <td>2017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>0.304316</td>\n",
       "      <td>United Arab Emirates</td>\n",
       "      <td>0.562150</td>\n",
       "      <td>1.573520</td>\n",
       "      <td>0.265910</td>\n",
       "      <td>27</td>\n",
       "      <td>6.573</td>\n",
       "      <td>0.729930</td>\n",
       "      <td>Middle East and Northern Africa</td>\n",
       "      <td>0.871140</td>\n",
       "      <td>2016</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>111</th>\n",
       "      <td>0.304316</td>\n",
       "      <td>Singapore</td>\n",
       "      <td>0.487700</td>\n",
       "      <td>1.645550</td>\n",
       "      <td>0.327060</td>\n",
       "      <td>21</td>\n",
       "      <td>6.739</td>\n",
       "      <td>0.947190</td>\n",
       "      <td>Southeastern Asia</td>\n",
       "      <td>0.867580</td>\n",
       "      <td>2016</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>288</th>\n",
       "      <td>0.039000</td>\n",
       "      <td>Croatia</td>\n",
       "      <td>0.380000</td>\n",
       "      <td>1.115000</td>\n",
       "      <td>0.120000</td>\n",
       "      <td>77</td>\n",
       "      <td>5.321</td>\n",
       "      <td>0.737000</td>\n",
       "      <td>Central and Eastern Europe</td>\n",
       "      <td>1.161000</td>\n",
       "      <td>2018</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>358</th>\n",
       "      <td>0.032000</td>\n",
       "      <td>Nigeria</td>\n",
       "      <td>0.462000</td>\n",
       "      <td>0.689000</td>\n",
       "      <td>0.201000</td>\n",
       "      <td>85</td>\n",
       "      <td>5.155</td>\n",
       "      <td>0.048000</td>\n",
       "      <td>Sub-Saharan Africa</td>\n",
       "      <td>1.172000</td>\n",
       "      <td>2018</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>0.182000</td>\n",
       "      <td>United Arab Emirates</td>\n",
       "      <td>0.598000</td>\n",
       "      <td>1.503000</td>\n",
       "      <td>0.262000</td>\n",
       "      <td>21</td>\n",
       "      <td>6.825</td>\n",
       "      <td>0.825000</td>\n",
       "      <td>Middle East and Northern Africa</td>\n",
       "      <td>1.310000</td>\n",
       "      <td>2019</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>417</th>\n",
       "      <td>0.099332</td>\n",
       "      <td>Philippines</td>\n",
       "      <td>0.585215</td>\n",
       "      <td>0.857699</td>\n",
       "      <td>0.193513</td>\n",
       "      <td>67</td>\n",
       "      <td>5.430</td>\n",
       "      <td>0.468009</td>\n",
       "      <td>Southeastern Asia</td>\n",
       "      <td>1.253918</td>\n",
       "      <td>2017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>304</th>\n",
       "      <td>0.028000</td>\n",
       "      <td>Jamaica</td>\n",
       "      <td>0.490000</td>\n",
       "      <td>0.831000</td>\n",
       "      <td>0.107000</td>\n",
       "      <td>54</td>\n",
       "      <td>5.890</td>\n",
       "      <td>0.831000</td>\n",
       "      <td>Latin America and Caribbean</td>\n",
       "      <td>1.478000</td>\n",
       "      <td>2019</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>237</th>\n",
       "      <td>0.010091</td>\n",
       "      <td>Moldova</td>\n",
       "      <td>0.240729</td>\n",
       "      <td>0.728871</td>\n",
       "      <td>0.208779</td>\n",
       "      <td>53</td>\n",
       "      <td>5.838</td>\n",
       "      <td>0.589465</td>\n",
       "      <td>Central and Eastern Europe</td>\n",
       "      <td>1.251826</td>\n",
       "      <td>2017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>324</th>\n",
       "      <td>0.028000</td>\n",
       "      <td>Turkmenistan</td>\n",
       "      <td>0.394000</td>\n",
       "      <td>1.052000</td>\n",
       "      <td>0.244000</td>\n",
       "      <td>82</td>\n",
       "      <td>5.247</td>\n",
       "      <td>0.657000</td>\n",
       "      <td>Central and Eastern Europe</td>\n",
       "      <td>1.538000</td>\n",
       "      <td>2019</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     corruption_perception               country   freedom  gdp_per_capita  \\\n",
       "262               0.130688             Nicaragua  0.447552        0.737299   \n",
       "96                0.304316  United Arab Emirates  0.562150        1.573520   \n",
       "111               0.304316             Singapore  0.487700        1.645550   \n",
       "288               0.039000               Croatia  0.380000        1.115000   \n",
       "358               0.032000               Nigeria  0.462000        0.689000   \n",
       "99                0.182000  United Arab Emirates  0.598000        1.503000   \n",
       "417               0.099332           Philippines  0.585215        0.857699   \n",
       "304               0.028000               Jamaica  0.490000        0.831000   \n",
       "237               0.010091               Moldova  0.240729        0.728871   \n",
       "324               0.028000          Turkmenistan  0.394000        1.052000   \n",
       "\n",
       "     generosity  happiness_rank  happiness_score  life_expectancy  \\\n",
       "262    0.301674              41            6.071         0.653096   \n",
       "96     0.265910              27            6.573         0.729930   \n",
       "111    0.327060              21            6.739         0.947190   \n",
       "288    0.120000              77            5.321         0.737000   \n",
       "358    0.201000              85            5.155         0.048000   \n",
       "99     0.262000              21            6.825         0.825000   \n",
       "417    0.193513              67            5.430         0.468009   \n",
       "304    0.107000              54            5.890         0.831000   \n",
       "237    0.208779              53            5.838         0.589465   \n",
       "324    0.244000              82            5.247         0.657000   \n",
       "\n",
       "                              region  social_support  year  \n",
       "262      Latin America and Caribbean        1.287216  2017  \n",
       "96   Middle East and Northern Africa        0.871140  2016  \n",
       "111                Southeastern Asia        0.867580  2016  \n",
       "288       Central and Eastern Europe        1.161000  2018  \n",
       "358               Sub-Saharan Africa        1.172000  2018  \n",
       "99   Middle East and Northern Africa        1.310000  2019  \n",
       "417                Southeastern Asia        1.253918  2017  \n",
       "304      Latin America and Caribbean        1.478000  2019  \n",
       "237       Central and Eastern Europe        1.251826  2017  \n",
       "324       Central and Eastern Europe        1.538000  2019  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_test = pd.read_csv('data/data_clean.csv')\n",
    "print(df_test.shape)\n",
    "display(df_test.sample(10))"
   ]
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}