models.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Loading data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/pg/091z68c93wb7x_j1473t961h0000gn/T/ipykernel_2965/4064310478.py:14: FutureWarning: Setting an item of incompatible dtype is deprecated and will raise an error in a future version of pandas. Value '[ 39985.17595884  46647.32056942  45295.64594906  37496.66241938\n",
      " 503521.1748     461001.608928   363654.1818     335680.7832\n",
      "  51354.68408582 290923.34544    279733.986      212821.6165488\n",
      " 212597.82936    210359.957472   183505.494816   131031.87478618\n",
      " 167840.3916     154804.7878524  154413.160272   151056.35244\n",
      " 146356.8214752  165894.56199338 139866.993      134272.31328\n",
      " 132034.441392   128677.63356    123082.95384    118607.210064\n",
      " 117488.27412    117488.27412    115250.402232   111893.5944\n",
      " 127611.28760537 111893.5944     109655.722512   106298.91468\n",
      " 101823.170904   126296.53787117  97347.427128   121990.87235866\n",
      "  95109.55524    102088.35872273  80563.387968    78325.51608\n",
      "  89328.01321735  89328.01321735  82946.72152872  78481.04817622\n",
      "  50075.74030183  67136.15664     76566.54877603  86111.07237835\n",
      "  67136.15664     62660.412864    70186.37602334  70186.37602334\n",
      "  66356.25828703  61252.79144645  51202.50879744  57424.91158202\n",
      "  56148.20566992  55127.73608899  53595.91278166  46995.309648\n",
      "  53595.91278166  52320.3258055   51043.61989339  47215.74002897\n",
      "  44663.4471407   43387.86016454  25521.8099467   22378.71888\n",
      "  24246.22297054  16784.03916     17864.9312819 ]' has dtype incompatible with int64, please explicitly cast to a compatible dtype first.\n",
      "  data.loc[data['work_year'] == 2020, 'salary_in_usd'] *= np.prod(inflation[1:4])\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>work_year</th>\n",
       "      <th>experience_level</th>\n",
       "      <th>employment_type</th>\n",
       "      <th>job_title</th>\n",
       "      <th>salary</th>\n",
       "      <th>salary_currency</th>\n",
       "      <th>salary_in_usd</th>\n",
       "      <th>employee_residence</th>\n",
       "      <th>remote_ratio</th>\n",
       "      <th>company_location</th>\n",
       "      <th>company_size</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2021</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>Data Scientist</td>\n",
       "      <td>30400000</td>\n",
       "      <td>CLP</td>\n",
       "      <td>44051.0</td>\n",
       "      <td>CL</td>\n",
       "      <td>100</td>\n",
       "      <td>CL</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2021</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>BI Data Analyst</td>\n",
       "      <td>11000000</td>\n",
       "      <td>HUF</td>\n",
       "      <td>39893.0</td>\n",
       "      <td>HU</td>\n",
       "      <td>50</td>\n",
       "      <td>US</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2020</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>Data Scientist</td>\n",
       "      <td>11000000</td>\n",
       "      <td>HUF</td>\n",
       "      <td>39985.0</td>\n",
       "      <td>HU</td>\n",
       "      <td>50</td>\n",
       "      <td>HU</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2021</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>ML Engineer</td>\n",
       "      <td>8500000</td>\n",
       "      <td>JPY</td>\n",
       "      <td>85118.0</td>\n",
       "      <td>JP</td>\n",
       "      <td>50</td>\n",
       "      <td>JP</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2022</td>\n",
       "      <td>SE</td>\n",
       "      <td>FT</td>\n",
       "      <td>Lead Machine Learning Engineer</td>\n",
       "      <td>7500000</td>\n",
       "      <td>INR</td>\n",
       "      <td>101300.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>50</td>\n",
       "      <td>IN</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2021</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>ML Engineer</td>\n",
       "      <td>7000000</td>\n",
       "      <td>JPY</td>\n",
       "      <td>70097.0</td>\n",
       "      <td>JP</td>\n",
       "      <td>50</td>\n",
       "      <td>JP</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2021</td>\n",
       "      <td>SE</td>\n",
       "      <td>FT</td>\n",
       "      <td>Data Science Manager</td>\n",
       "      <td>7000000</td>\n",
       "      <td>INR</td>\n",
       "      <td>104153.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>50</td>\n",
       "      <td>IN</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>2022</td>\n",
       "      <td>EN</td>\n",
       "      <td>FT</td>\n",
       "      <td>Data Scientist</td>\n",
       "      <td>6600000</td>\n",
       "      <td>HUF</td>\n",
       "      <td>18780.0</td>\n",
       "      <td>HU</td>\n",
       "      <td>100</td>\n",
       "      <td>HU</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>2022</td>\n",
       "      <td>EX</td>\n",
       "      <td>FT</td>\n",
       "      <td>Head of Machine Learning</td>\n",
       "      <td>6000000</td>\n",
       "      <td>INR</td>\n",
       "      <td>81040.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>50</td>\n",
       "      <td>IN</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2022</td>\n",
       "      <td>EN</td>\n",
       "      <td>FT</td>\n",
       "      <td>Research Engineer</td>\n",
       "      <td>5500000</td>\n",
       "      <td>JPY</td>\n",
       "      <td>44401.0</td>\n",
       "      <td>JP</td>\n",
       "      <td>50</td>\n",
       "      <td>JP</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>2023</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>Head of Data Science</td>\n",
       "      <td>5000000</td>\n",
       "      <td>INR</td>\n",
       "      <td>63724.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>50</td>\n",
       "      <td>IN</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>2023</td>\n",
       "      <td>EN</td>\n",
       "      <td>FT</td>\n",
       "      <td>AI Programmer</td>\n",
       "      <td>4950806</td>\n",
       "      <td>INR</td>\n",
       "      <td>63097.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>0</td>\n",
       "      <td>IN</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>2021</td>\n",
       "      <td>SE</td>\n",
       "      <td>FT</td>\n",
       "      <td>Machine Learning Engineer</td>\n",
       "      <td>4900000</td>\n",
       "      <td>INR</td>\n",
       "      <td>72907.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>0</td>\n",
       "      <td>IN</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>2022</td>\n",
       "      <td>SE</td>\n",
       "      <td>FT</td>\n",
       "      <td>Lead Data Scientist</td>\n",
       "      <td>4460000</td>\n",
       "      <td>INR</td>\n",
       "      <td>60240.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>0</td>\n",
       "      <td>IN</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>2020</td>\n",
       "      <td>EN</td>\n",
       "      <td>FT</td>\n",
       "      <td>Data Engineer</td>\n",
       "      <td>4450000</td>\n",
       "      <td>JPY</td>\n",
       "      <td>46647.0</td>\n",
       "      <td>JP</td>\n",
       "      <td>100</td>\n",
       "      <td>JP</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>2022</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>Data Scientist</td>\n",
       "      <td>4200000</td>\n",
       "      <td>INR</td>\n",
       "      <td>56728.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>100</td>\n",
       "      <td>ID</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>2023</td>\n",
       "      <td>EN</td>\n",
       "      <td>FT</td>\n",
       "      <td>Applied Machine Learning Scientist</td>\n",
       "      <td>4000000</td>\n",
       "      <td>INR</td>\n",
       "      <td>50979.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>100</td>\n",
       "      <td>DE</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>2021</td>\n",
       "      <td>SE</td>\n",
       "      <td>FT</td>\n",
       "      <td>Data Scientist</td>\n",
       "      <td>4000000</td>\n",
       "      <td>INR</td>\n",
       "      <td>59516.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>100</td>\n",
       "      <td>IN</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>2021</td>\n",
       "      <td>SE</td>\n",
       "      <td>FT</td>\n",
       "      <td>Data Science Manager</td>\n",
       "      <td>4000000</td>\n",
       "      <td>INR</td>\n",
       "      <td>59516.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>50</td>\n",
       "      <td>US</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>2023</td>\n",
       "      <td>SE</td>\n",
       "      <td>FT</td>\n",
       "      <td>Data Science Manager</td>\n",
       "      <td>3600000</td>\n",
       "      <td>INR</td>\n",
       "      <td>45880.0</td>\n",
       "      <td>IN</td>\n",
       "      <td>100</td>\n",
       "      <td>IN</td>\n",
       "      <td>L</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    work_year experience_level employment_type  \\\n",
       "0        2021               MI              FT   \n",
       "1        2021               MI              FT   \n",
       "2        2020               MI              FT   \n",
       "3        2021               MI              FT   \n",
       "4        2022               SE              FT   \n",
       "5        2021               MI              FT   \n",
       "6        2021               SE              FT   \n",
       "7        2022               EN              FT   \n",
       "8        2022               EX              FT   \n",
       "9        2022               EN              FT   \n",
       "10       2023               MI              FT   \n",
       "11       2023               EN              FT   \n",
       "12       2021               SE              FT   \n",
       "13       2022               SE              FT   \n",
       "14       2020               EN              FT   \n",
       "15       2022               MI              FT   \n",
       "16       2023               EN              FT   \n",
       "17       2021               SE              FT   \n",
       "18       2021               SE              FT   \n",
       "19       2023               SE              FT   \n",
       "\n",
       "                             job_title    salary salary_currency  \\\n",
       "0                       Data Scientist  30400000             CLP   \n",
       "1                      BI Data Analyst  11000000             HUF   \n",
       "2                       Data Scientist  11000000             HUF   \n",
       "3                          ML Engineer   8500000             JPY   \n",
       "4       Lead Machine Learning Engineer   7500000             INR   \n",
       "5                          ML Engineer   7000000             JPY   \n",
       "6                 Data Science Manager   7000000             INR   \n",
       "7                       Data Scientist   6600000             HUF   \n",
       "8             Head of Machine Learning   6000000             INR   \n",
       "9                    Research Engineer   5500000             JPY   \n",
       "10                Head of Data Science   5000000             INR   \n",
       "11                       AI Programmer   4950806             INR   \n",
       "12           Machine Learning Engineer   4900000             INR   \n",
       "13                 Lead Data Scientist   4460000             INR   \n",
       "14                       Data Engineer   4450000             JPY   \n",
       "15                      Data Scientist   4200000             INR   \n",
       "16  Applied Machine Learning Scientist   4000000             INR   \n",
       "17                      Data Scientist   4000000             INR   \n",
       "18                Data Science Manager   4000000             INR   \n",
       "19                Data Science Manager   3600000             INR   \n",
       "\n",
       "    salary_in_usd employee_residence  remote_ratio company_location  \\\n",
       "0         44051.0                 CL           100               CL   \n",
       "1         39893.0                 HU            50               US   \n",
       "2         39985.0                 HU            50               HU   \n",
       "3         85118.0                 JP            50               JP   \n",
       "4        101300.0                 IN            50               IN   \n",
       "5         70097.0                 JP            50               JP   \n",
       "6        104153.0                 IN            50               IN   \n",
       "7         18780.0                 HU           100               HU   \n",
       "8         81040.0                 IN            50               IN   \n",
       "9         44401.0                 JP            50               JP   \n",
       "10        63724.0                 IN            50               IN   \n",
       "11        63097.0                 IN             0               IN   \n",
       "12        72907.0                 IN             0               IN   \n",
       "13        60240.0                 IN             0               IN   \n",
       "14        46647.0                 JP           100               JP   \n",
       "15        56728.0                 IN           100               ID   \n",
       "16        50979.0                 IN           100               DE   \n",
       "17        59516.0                 IN           100               IN   \n",
       "18        59516.0                 IN            50               US   \n",
       "19        45880.0                 IN           100               IN   \n",
       "\n",
       "   company_size  \n",
       "0             L  \n",
       "1             L  \n",
       "2             L  \n",
       "3             S  \n",
       "4             L  \n",
       "5             S  \n",
       "6             L  \n",
       "7             M  \n",
       "8             L  \n",
       "9             L  \n",
       "10            L  \n",
       "11            S  \n",
       "12            L  \n",
       "13            L  \n",
       "14            S  \n",
       "15            L  \n",
       "16            L  \n",
       "17            L  \n",
       "18            L  \n",
       "19            L  "
      ]
     },
     "execution_count": 144,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import sklearn\n",
    "\n",
    "data = pd.read_csv(\"DataScience_salaries_2024.csv\")\n",
    "\n",
    "# Yearly Inflation Rates\n",
    "inflation = [1.022, 1.017, 1.036, 1.062, 1.048]\n",
    "\n",
    "# Calculate the inflation adjusted salary for each year\n",
    "data.loc[data['work_year'] == 2019, 'salary_in_usd'] *= np.prod(inflation[0:4])\n",
    "data.loc[data['work_year'] == 2020, 'salary_in_usd'] *= np.prod(inflation[1:4])\n",
    "data.loc[data['work_year'] == 2021, 'salary_in_usd'] *= np.prod(inflation[2:4])\n",
    "data.loc[data['work_year'] == 2022, 'salary_in_usd'] *= np.prod(inflation[3:4])\n",
    "data.loc[data['work_year'] == 2023, 'salary_in_usd'] *= inflation[4]\n",
    "\n",
    "data['salary_in_usd'] = data['salary_in_usd'].apply(np.round)\n",
    "\n",
    "data.head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Basic information"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Instances: 14838\n",
      "Attributes: 11\n",
      "\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 14838 entries, 0 to 14837\n",
      "Data columns (total 11 columns):\n",
      " #   Column              Non-Null Count  Dtype  \n",
      "---  ------              --------------  -----  \n",
      " 0   work_year           14838 non-null  int64  \n",
      " 1   experience_level    14838 non-null  object \n",
      " 2   employment_type     14838 non-null  object \n",
      " 3   job_title           14838 non-null  object \n",
      " 4   salary              14838 non-null  int64  \n",
      " 5   salary_currency     14838 non-null  object \n",
      " 6   salary_in_usd       14838 non-null  float64\n",
      " 7   employee_residence  14838 non-null  object \n",
      " 8   remote_ratio        14838 non-null  int64  \n",
      " 9   company_location    14838 non-null  object \n",
      " 10  company_size        14838 non-null  object \n",
      "dtypes: float64(1), int64(3), object(7)\n",
      "memory usage: 1.2+ MB\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>work_year</th>\n",
       "      <th>salary</th>\n",
       "      <th>salary_in_usd</th>\n",
       "      <th>remote_ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>14838.000000</td>\n",
       "      <td>1.483800e+04</td>\n",
       "      <td>14838.000000</td>\n",
       "      <td>14838.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>2023.138900</td>\n",
       "      <td>1.650227e+05</td>\n",
       "      <td>155247.734398</td>\n",
       "      <td>32.760480</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.700799</td>\n",
       "      <td>3.562354e+05</td>\n",
       "      <td>71197.357472</td>\n",
       "      <td>46.488278</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2020.000000</td>\n",
       "      <td>1.400000e+04</td>\n",
       "      <td>15930.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2023.000000</td>\n",
       "      <td>1.021000e+05</td>\n",
       "      <td>105700.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2023.000000</td>\n",
       "      <td>1.422000e+05</td>\n",
       "      <td>146720.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2024.000000</td>\n",
       "      <td>1.875000e+05</td>\n",
       "      <td>193880.000000</td>\n",
       "      <td>100.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2024.000000</td>\n",
       "      <td>3.040000e+07</td>\n",
       "      <td>800000.000000</td>\n",
       "      <td>100.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          work_year        salary  salary_in_usd  remote_ratio\n",
       "count  14838.000000  1.483800e+04   14838.000000  14838.000000\n",
       "mean    2023.138900  1.650227e+05  155247.734398     32.760480\n",
       "std        0.700799  3.562354e+05   71197.357472     46.488278\n",
       "min     2020.000000  1.400000e+04   15930.000000      0.000000\n",
       "25%     2023.000000  1.021000e+05  105700.000000      0.000000\n",
       "50%     2023.000000  1.422000e+05  146720.000000      0.000000\n",
       "75%     2024.000000  1.875000e+05  193880.000000    100.000000\n",
       "max     2024.000000  3.040000e+07  800000.000000    100.000000"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check for missing values and printing info about the dataset\n",
    "print(\"Instances: \" + str(data.shape[0]))\n",
    "print(\"Attributes: \" + str(data.shape[1]) +\"\\n\")\n",
    "data.info()\n",
    "data.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are 14838 instance and 11 attributes. There are 3 numerical attributes and 8 categorical attributes. There is no missing value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "       salary_in_usd employee_residence company_location  remote_ratio\n",
      "0            44051.0                 CL               CL           100\n",
      "1            39893.0                 HU               US            50\n",
      "2            39985.0                 HU               HU            50\n",
      "3            85118.0                 JP               JP            50\n",
      "4           101300.0                 IN               IN            50\n",
      "...              ...                ...              ...           ...\n",
      "14833        15930.0                 GH               GH           100\n",
      "14834        16784.0                 NG               CA             0\n",
      "14835        16503.0                 TH               TH           100\n",
      "14836        15930.0                 ID               ID             0\n",
      "14837        17865.0                 DE               DE           100\n",
      "\n",
      "[14838 rows x 4 columns]\n"
     ]
    }
   ],
   "source": [
    "# Defining new dataFrame with already numerical columns\n",
    "\n",
    "encoded_data = data[['salary_in_usd','employee_residence','company_location','remote_ratio']].copy()\n",
    "\n",
    "print(encoded_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Using IQR to remove outliers\n",
    "\n",
    "Q1 = data['salary_in_usd'].quantile(0.25)\n",
    "Q3 = data['salary_in_usd'].quantile(0.75)\n",
    "IQR = Q3 - Q1\n",
    "lower_bound = Q1 - 1.5 * IQR\n",
    "upper_bound = Q3 + 1.5 * IQR\n",
    "\n",
    "data = data[(data['salary_in_usd'] >= lower_bound) & (data['salary_in_usd'] <= upper_bound)].reset_index().copy()\n",
    "data.drop('index', axis = 1, inplace = True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function for Job Title Preprocessing\n",
    "def process_job_title(job_titles):\n",
    "  #define keywords\n",
    "  job_cat_key = ['Data', 'Machine Learning', 'Business Intelligence', 'Artificial Intelligence', 'Research', 'Other']\n",
    "  job_title_key = ['Engineer', 'Analyst', 'Developer', 'Scientist', 'Speicialist', 'Manager']\n",
    "\n",
    "  job_cat_new = []\n",
    "  job_title_new = []\n",
    "\n",
    "  def check_job_cat_or_type(jobs, title):\n",
    "    for job in jobs:\n",
    "       if job in title:\n",
    "          return job\n",
    "    return None\n",
    "\n",
    "  for i, title in enumerate(job_titles):\n",
    "     #replace abbreviations\n",
    "     title = title.replace('ML', 'Machine Learning').replace('BI', 'Business Intelligence').replace('AI', 'Artificial Intelligence')\n",
    "\n",
    "  for i, title in enumerate(job_titles):\n",
    "    check1 = check_job_cat_or_type(job_cat_key, title)\n",
    "    check2 = check_job_cat_or_type(job_title_key, title)\n",
    "\n",
    "    if check1 is not None:\n",
    "        job_cat_new.append(check1)\n",
    "    else:\n",
    "        job_cat_new.append('Other')\n",
    "    \n",
    "    if check2 is not None:\n",
    "        job_title_new.append(check2)\n",
    "    else:\n",
    "        job_title_new.append('Other')\n",
    "\n",
    "  return job_cat_new, job_title_new\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function for Location Preprocessing\n",
    "import pycountry\n",
    "import pycountry_convert as pc\n",
    "\n",
    "def process_location(job_data):\n",
    "    location_dic = job_data.value_counts().to_dict()\n",
    "    filter_threshold = 0.6\n",
    "\n",
    "    job_data = job_data.values\n",
    "    \n",
    "    def get_continent(country_code):\n",
    "        country_name = pycountry.countries.get(alpha_2 = country_code)\n",
    "        if country_name:\n",
    "            return pc.country_alpha2_to_continent_code(country_code)\n",
    "        else:\n",
    "            return 'Other'\n",
    "\n",
    "    for i in range(len(job_data)):\n",
    "        if location_dic[job_data[i]] / len(job_data) < filter_threshold:\n",
    "            job_data[i] = get_continent(job_data[i])\n",
    "    \n",
    "    return job_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "employee_residence\n",
      "US    12696\n",
      "EU     1176\n",
      "NA      408\n",
      "AS      149\n",
      "OC       57\n",
      "SA       56\n",
      "AF       54\n",
      "Name: count, dtype: int64\n",
      "company_location\n",
      "US    12744\n",
      "EU     1172\n",
      "NA      409\n",
      "AS      117\n",
      "OC       62\n",
      "SA       46\n",
      "AF       46\n",
      "Name: count, dtype: int64\n",
      "job_category\n",
      "Data                     10115\n",
      "Machine Learning          1735\n",
      "Other                     1430\n",
      "Research                   816\n",
      "Business Intelligence      500\n",
      "Name: count, dtype: int64\n",
      "job_title\n",
      "Engineer     6099\n",
      "Scientist    4042\n",
      "Analyst      2640\n",
      "Other        1173\n",
      "Manager       418\n",
      "Developer     224\n",
      "Name: count, dtype: int64\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>work_year</th>\n",
       "      <th>experience_level</th>\n",
       "      <th>employment_type</th>\n",
       "      <th>job_title</th>\n",
       "      <th>salary</th>\n",
       "      <th>salary_currency</th>\n",
       "      <th>salary_in_usd</th>\n",
       "      <th>employee_residence</th>\n",
       "      <th>remote_ratio</th>\n",
       "      <th>company_location</th>\n",
       "      <th>company_size</th>\n",
       "      <th>job_category</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2021</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>Scientist</td>\n",
       "      <td>30400000</td>\n",
       "      <td>CLP</td>\n",
       "      <td>44051.0</td>\n",
       "      <td>SA</td>\n",
       "      <td>100</td>\n",
       "      <td>SA</td>\n",
       "      <td>L</td>\n",
       "      <td>Data</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2021</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>Analyst</td>\n",
       "      <td>11000000</td>\n",
       "      <td>HUF</td>\n",
       "      <td>39893.0</td>\n",
       "      <td>EU</td>\n",
       "      <td>50</td>\n",
       "      <td>US</td>\n",
       "      <td>L</td>\n",
       "      <td>Data</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2020</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>Scientist</td>\n",
       "      <td>11000000</td>\n",
       "      <td>HUF</td>\n",
       "      <td>39985.0</td>\n",
       "      <td>EU</td>\n",
       "      <td>50</td>\n",
       "      <td>EU</td>\n",
       "      <td>L</td>\n",
       "      <td>Data</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2021</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>Engineer</td>\n",
       "      <td>8500000</td>\n",
       "      <td>JPY</td>\n",
       "      <td>85118.0</td>\n",
       "      <td>AS</td>\n",
       "      <td>50</td>\n",
       "      <td>AS</td>\n",
       "      <td>S</td>\n",
       "      <td>Other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2022</td>\n",
       "      <td>SE</td>\n",
       "      <td>FT</td>\n",
       "      <td>Engineer</td>\n",
       "      <td>7500000</td>\n",
       "      <td>INR</td>\n",
       "      <td>101300.0</td>\n",
       "      <td>AS</td>\n",
       "      <td>50</td>\n",
       "      <td>AS</td>\n",
       "      <td>L</td>\n",
       "      <td>Machine Learning</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14591</th>\n",
       "      <td>2022</td>\n",
       "      <td>MI</td>\n",
       "      <td>FT</td>\n",
       "      <td>Developer</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>15930.0</td>\n",
       "      <td>AF</td>\n",
       "      <td>100</td>\n",
       "      <td>AF</td>\n",
       "      <td>M</td>\n",
       "      <td>Business Intelligence</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14592</th>\n",
       "      <td>2020</td>\n",
       "      <td>EX</td>\n",
       "      <td>FT</td>\n",
       "      <td>Analyst</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>16784.0</td>\n",
       "      <td>AF</td>\n",
       "      <td>0</td>\n",
       "      <td>NA</td>\n",
       "      <td>M</td>\n",
       "      <td>Data</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14593</th>\n",
       "      <td>2021</td>\n",
       "      <td>EN</td>\n",
       "      <td>FT</td>\n",
       "      <td>Developer</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>16503.0</td>\n",
       "      <td>AS</td>\n",
       "      <td>100</td>\n",
       "      <td>AS</td>\n",
       "      <td>L</td>\n",
       "      <td>Machine Learning</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14594</th>\n",
       "      <td>2022</td>\n",
       "      <td>EN</td>\n",
       "      <td>FT</td>\n",
       "      <td>Analyst</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>15930.0</td>\n",
       "      <td>AS</td>\n",
       "      <td>0</td>\n",
       "      <td>AS</td>\n",
       "      <td>L</td>\n",
       "      <td>Data</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14595</th>\n",
       "      <td>2020</td>\n",
       "      <td>EN</td>\n",
       "      <td>PT</td>\n",
       "      <td>Engineer</td>\n",
       "      <td>14000</td>\n",
       "      <td>EUR</td>\n",
       "      <td>17865.0</td>\n",
       "      <td>EU</td>\n",
       "      <td>100</td>\n",
       "      <td>EU</td>\n",
       "      <td>S</td>\n",
       "      <td>Other</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>14596 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       work_year experience_level employment_type  job_title    salary  \\\n",
       "0           2021               MI              FT  Scientist  30400000   \n",
       "1           2021               MI              FT    Analyst  11000000   \n",
       "2           2020               MI              FT  Scientist  11000000   \n",
       "3           2021               MI              FT   Engineer   8500000   \n",
       "4           2022               SE              FT   Engineer   7500000   \n",
       "...          ...              ...             ...        ...       ...   \n",
       "14591       2022               MI              FT  Developer     15000   \n",
       "14592       2020               EX              FT    Analyst     15000   \n",
       "14593       2021               EN              FT  Developer     15000   \n",
       "14594       2022               EN              FT    Analyst     15000   \n",
       "14595       2020               EN              PT   Engineer     14000   \n",
       "\n",
       "      salary_currency  salary_in_usd employee_residence  remote_ratio  \\\n",
       "0                 CLP        44051.0                 SA           100   \n",
       "1                 HUF        39893.0                 EU            50   \n",
       "2                 HUF        39985.0                 EU            50   \n",
       "3                 JPY        85118.0                 AS            50   \n",
       "4                 INR       101300.0                 AS            50   \n",
       "...               ...            ...                ...           ...   \n",
       "14591             USD        15930.0                 AF           100   \n",
       "14592             USD        16784.0                 AF             0   \n",
       "14593             USD        16503.0                 AS           100   \n",
       "14594             USD        15930.0                 AS             0   \n",
       "14595             EUR        17865.0                 EU           100   \n",
       "\n",
       "      company_location company_size           job_category  \n",
       "0                   SA            L                   Data  \n",
       "1                   US            L                   Data  \n",
       "2                   EU            L                   Data  \n",
       "3                   AS            S                  Other  \n",
       "4                   AS            L       Machine Learning  \n",
       "...                ...          ...                    ...  \n",
       "14591               AF            M  Business Intelligence  \n",
       "14592               NA            M                   Data  \n",
       "14593               AS            L       Machine Learning  \n",
       "14594               AS            L                   Data  \n",
       "14595               EU            S                  Other  \n",
       "\n",
       "[14596 rows x 12 columns]"
      ]
     },
     "execution_count": 152,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Running the preprocessing functions and adding them to the data\n",
    "\n",
    "data['employee_residence'] = process_location(data['employee_residence'])\n",
    "data['company_location'] = process_location(data['company_location'])\n",
    "\n",
    "print(data['employee_residence'].value_counts())\n",
    "print(data['company_location'].value_counts())\n",
    "\n",
    "data['job_category'], data['job_title'] = process_job_title(data['job_title'])\n",
    "\n",
    "print(data['job_category'].value_counts())\n",
    "print(data['job_title'].value_counts())\n",
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "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>work_year</th>\n",
       "      <th>salary</th>\n",
       "      <th>salary_currency</th>\n",
       "      <th>salary_in_usd</th>\n",
       "      <th>ode1__experience_level</th>\n",
       "      <th>ohe__employee_residence_AF</th>\n",
       "      <th>ohe__employee_residence_AS</th>\n",
       "      <th>ohe__employee_residence_EU</th>\n",
       "      <th>ohe__employee_residence_NA</th>\n",
       "      <th>ohe__employee_residence_OC</th>\n",
       "      <th>...</th>\n",
       "      <th>ohe__employment_type_FL</th>\n",
       "      <th>ohe__employment_type_FT</th>\n",
       "      <th>ohe__employment_type_PT</th>\n",
       "      <th>ohe__company_size_L</th>\n",
       "      <th>ohe__company_size_M</th>\n",
       "      <th>ohe__company_size_S</th>\n",
       "      <th>remainder__work_year</th>\n",
       "      <th>remainder__salary</th>\n",
       "      <th>remainder__salary_currency</th>\n",
       "      <th>remainder__salary_in_usd</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2021</td>\n",
       "      <td>30400000</td>\n",
       "      <td>CLP</td>\n",
       "      <td>44051.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>30400000</td>\n",
       "      <td>CLP</td>\n",
       "      <td>44051.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2021</td>\n",
       "      <td>11000000</td>\n",
       "      <td>HUF</td>\n",
       "      <td>39893.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>11000000</td>\n",
       "      <td>HUF</td>\n",
       "      <td>39893.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2020</td>\n",
       "      <td>11000000</td>\n",
       "      <td>HUF</td>\n",
       "      <td>39985.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2020</td>\n",
       "      <td>11000000</td>\n",
       "      <td>HUF</td>\n",
       "      <td>39985.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2021</td>\n",
       "      <td>8500000</td>\n",
       "      <td>JPY</td>\n",
       "      <td>85118.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>8500000</td>\n",
       "      <td>JPY</td>\n",
       "      <td>85118.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2022</td>\n",
       "      <td>7500000</td>\n",
       "      <td>INR</td>\n",
       "      <td>101300.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2022</td>\n",
       "      <td>7500000</td>\n",
       "      <td>INR</td>\n",
       "      <td>101300.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14591</th>\n",
       "      <td>2022</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>15930.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2022</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>15930.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14592</th>\n",
       "      <td>2020</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>16784.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2020</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>16784.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14593</th>\n",
       "      <td>2021</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>16503.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2021</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>16503.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14594</th>\n",
       "      <td>2022</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>15930.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2022</td>\n",
       "      <td>15000</td>\n",
       "      <td>USD</td>\n",
       "      <td>15930.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14595</th>\n",
       "      <td>2020</td>\n",
       "      <td>14000</td>\n",
       "      <td>EUR</td>\n",
       "      <td>17865.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2020</td>\n",
       "      <td>14000</td>\n",
       "      <td>EUR</td>\n",
       "      <td>17865.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>14596 rows × 44 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       work_year    salary salary_currency  salary_in_usd  \\\n",
       "0           2021  30400000             CLP        44051.0   \n",
       "1           2021  11000000             HUF        39893.0   \n",
       "2           2020  11000000             HUF        39985.0   \n",
       "3           2021   8500000             JPY        85118.0   \n",
       "4           2022   7500000             INR       101300.0   \n",
       "...          ...       ...             ...            ...   \n",
       "14591       2022     15000             USD        15930.0   \n",
       "14592       2020     15000             USD        16784.0   \n",
       "14593       2021     15000             USD        16503.0   \n",
       "14594       2022     15000             USD        15930.0   \n",
       "14595       2020     14000             EUR        17865.0   \n",
       "\n",
       "      ode1__experience_level ohe__employee_residence_AF  \\\n",
       "0                        1.0                        0.0   \n",
       "1                        1.0                        0.0   \n",
       "2                        1.0                        0.0   \n",
       "3                        1.0                        0.0   \n",
       "4                        2.0                        0.0   \n",
       "...                      ...                        ...   \n",
       "14591                    1.0                        1.0   \n",
       "14592                    3.0                        1.0   \n",
       "14593                    0.0                        0.0   \n",
       "14594                    0.0                        0.0   \n",
       "14595                    0.0                        0.0   \n",
       "\n",
       "      ohe__employee_residence_AS ohe__employee_residence_EU  \\\n",
       "0                            0.0                        0.0   \n",
       "1                            0.0                        1.0   \n",
       "2                            0.0                        1.0   \n",
       "3                            1.0                        0.0   \n",
       "4                            1.0                        0.0   \n",
       "...                          ...                        ...   \n",
       "14591                        0.0                        0.0   \n",
       "14592                        0.0                        0.0   \n",
       "14593                        1.0                        0.0   \n",
       "14594                        1.0                        0.0   \n",
       "14595                        0.0                        1.0   \n",
       "\n",
       "      ohe__employee_residence_NA ohe__employee_residence_OC  ...  \\\n",
       "0                            0.0                        0.0  ...   \n",
       "1                            0.0                        0.0  ...   \n",
       "2                            0.0                        0.0  ...   \n",
       "3                            0.0                        0.0  ...   \n",
       "4                            0.0                        0.0  ...   \n",
       "...                          ...                        ...  ...   \n",
       "14591                        0.0                        0.0  ...   \n",
       "14592                        0.0                        0.0  ...   \n",
       "14593                        0.0                        0.0  ...   \n",
       "14594                        0.0                        0.0  ...   \n",
       "14595                        0.0                        0.0  ...   \n",
       "\n",
       "      ohe__employment_type_FL ohe__employment_type_FT ohe__employment_type_PT  \\\n",
       "0                         0.0                     1.0                     0.0   \n",
       "1                         0.0                     1.0                     0.0   \n",
       "2                         0.0                     1.0                     0.0   \n",
       "3                         0.0                     1.0                     0.0   \n",
       "4                         0.0                     1.0                     0.0   \n",
       "...                       ...                     ...                     ...   \n",
       "14591                     0.0                     1.0                     0.0   \n",
       "14592                     0.0                     1.0                     0.0   \n",
       "14593                     0.0                     1.0                     0.0   \n",
       "14594                     0.0                     1.0                     0.0   \n",
       "14595                     0.0                     0.0                     1.0   \n",
       "\n",
       "      ohe__company_size_L ohe__company_size_M ohe__company_size_S  \\\n",
       "0                     1.0                 0.0                 0.0   \n",
       "1                     1.0                 0.0                 0.0   \n",
       "2                     1.0                 0.0                 0.0   \n",
       "3                     0.0                 0.0                 1.0   \n",
       "4                     1.0                 0.0                 0.0   \n",
       "...                   ...                 ...                 ...   \n",
       "14591                 0.0                 1.0                 0.0   \n",
       "14592                 0.0                 1.0                 0.0   \n",
       "14593                 1.0                 0.0                 0.0   \n",
       "14594                 1.0                 0.0                 0.0   \n",
       "14595                 0.0                 0.0                 1.0   \n",
       "\n",
       "      remainder__work_year remainder__salary remainder__salary_currency  \\\n",
       "0                     2021          30400000                        CLP   \n",
       "1                     2021          11000000                        HUF   \n",
       "2                     2020          11000000                        HUF   \n",
       "3                     2021           8500000                        JPY   \n",
       "4                     2022           7500000                        INR   \n",
       "...                    ...               ...                        ...   \n",
       "14591                 2022             15000                        USD   \n",
       "14592                 2020             15000                        USD   \n",
       "14593                 2021             15000                        USD   \n",
       "14594                 2022             15000                        USD   \n",
       "14595                 2020             14000                        EUR   \n",
       "\n",
       "      remainder__salary_in_usd  \n",
       "0                      44051.0  \n",
       "1                      39893.0  \n",
       "2                      39985.0  \n",
       "3                      85118.0  \n",
       "4                     101300.0  \n",
       "...                        ...  \n",
       "14591                  15930.0  \n",
       "14592                  16784.0  \n",
       "14593                  16503.0  \n",
       "14594                  15930.0  \n",
       "14595                  17865.0  \n",
       "\n",
       "[14596 rows x 44 columns]"
      ]
     },
     "execution_count": 153,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import OneHotEncoder,OrdinalEncoder\n",
    "from sklearn.compose import ColumnTransformer\n",
    "\n",
    "one_hot_encoding_cat = ['employee_residence','company_location','remote_ratio', 'job_category', 'job_title', 'employment_type','company_size']\n",
    "ordinal_encoding_cat = ['experience_level']\n",
    "\n",
    "ode1 = OrdinalEncoder(categories=[['EN', 'MI', 'SE', 'EX']])\n",
    "ohe = OneHotEncoder(sparse_output = False)\n",
    "\n",
    "# columnTransformer to run the encodings at once\n",
    "ct = ColumnTransformer(\n",
    "    transformers=[\n",
    "        ('ode1', ode1, ordinal_encoding_cat),\n",
    "        ('ohe', ohe, one_hot_encoding_cat),\n",
    "    ],\n",
    "    remainder='passthrough'\n",
    ")\n",
    "\n",
    "trf = ct.fit_transform(data)\n",
    "\n",
    "# Adding the newly encoded columns to the dataset\n",
    "\n",
    "df = pd.DataFrame(trf,columns=ct.get_feature_names_out())\n",
    "\n",
    "# Dropping the old columns\n",
    "data.drop(one_hot_encoding_cat + ordinal_encoding_cat, axis=1, inplace=True)\n",
    "\n",
    "data = data.join(df)\n",
    "\n",
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Split Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "# Drop the columns that will artitically influence models\n",
    "\n",
    "data_drop = ['work_year', 'salary', 'salary_currency', 'remainder__salary_currency', 'remainder__salary_in_usd','remainder__salary','remainder__work_year']\n",
    "data.drop(data_drop, axis=1, inplace=True)\n",
    "\n",
    "# Standardize the salary_in_usd column\n",
    "scaler = StandardScaler()\n",
    "data['salary_in_usd'] = pd.DataFrame(scaler.fit_transform(data[['salary_in_usd']]))\n",
    "\n",
    "# Split the data into training and testing sets\n",
    "y = data['salary_in_usd']\n",
    "X = data.drop('salary_in_usd', axis=1, inplace=False)\n",
    "\n",
    "X_train_temp, X_test, y_train_temp, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n",
    "X_train, X_val, y_train, y_val = train_test_split(X_train_temp, y_train_temp, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "\n",
    "# Custom function to neatly calculate and display the metrics\n",
    "def calculate_metrics(y_test, y_pred, threshold): \n",
    "\n",
    "    # Compute mse, rmse, r2 score\n",
    "    mse = mean_squared_error(y_test, y_pred)\n",
    "    rmse = np.sqrt(mse)\n",
    "    r2score = r2_score(y_test, y_pred)\n",
    "\n",
    "    print(\"Mean Squared Error: \", mse)\n",
    "    print(\"Root Mean Squared Error: \", rmse)\n",
    "    print(\"R2 Score: \", r2score)\n",
    "\n",
    "    print(\"The accuracy of our model (based on R2 score) is {}%\".format(round(r2score, 2) * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Baseline Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Squared Error:  0.9875609519048835\n",
      "Root Mean Squared Error:  0.9937610134760185\n",
      "R2 Score:  -0.00010097444762435792\n",
      "The accuracy of our model (based on R2 score) is -0.0%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.dummy import DummyRegressor\n",
    "\n",
    "dummy = DummyRegressor(strategy = 'mean')\n",
    "\n",
    "dummy.fit(X_train, y_train)\n",
    "y_pred_dummy = dummy.predict(X_val)\n",
    "\n",
    "calculate_metrics(y_val, y_pred_dummy, 0.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Linear regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Squared Error:  0.6490004075221756\n",
      "Root Mean Squared Error:  0.8056056153740337\n",
      "R2 Score:  0.34275860266866076\n",
      "The accuracy of our model (based on R2 score) is 34.0%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LinearRegression\n",
    "\n",
    "# Create and train the linear regression model\n",
    "model = LinearRegression()\n",
    "model.fit(X_train, y_train)\n",
    "\n",
    "# Make predictions on the test set\n",
    "y_pred_linear = model.predict(X_val)\n",
    "\n",
    "calculate_metrics(y_val, y_pred_linear, 0.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Polynomial Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Squared Error:  1.1324528452112517e+23\n",
      "Root Mean Squared Error:  336519367230.36487\n",
      "R2 Score:  -1.1468327011382688e+23\n",
      "The accuracy of our model (based on R2 score) is -1.1468327011382687e+25%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import PolynomialFeatures\n",
    "\n",
    "poly = PolynomialFeatures(degree=2)\n",
    "X_train_poly = poly.fit_transform(X_train)\n",
    "X_test_poly = poly.transform(X_val)\n",
    "\n",
    "model_poly = LinearRegression()\n",
    "model_poly.fit(X_train_poly, y_train)\n",
    "\n",
    "y_pred_poly = model_poly.predict(X_test_poly)\n",
    "\n",
    "calculate_metrics(y_val, y_pred_poly, 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Squared Error:  0.6343608808900165\n",
      "Root Mean Squared Error:  0.7964677525738355\n",
      "R2 Score:  0.3575840216182795\n",
      "The accuracy of our model (based on R2 score) is 36.0%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.svm import SVR\n",
    "\n",
    "model_svr = SVR()\n",
    "model_svr.fit(X_train, y_train)\n",
    "\n",
    "y_pred_svr = model_svr.predict(X_val)\n",
    "\n",
    "calculate_metrics(y_val, y_pred_svr, 0.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Decision Tree Regressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Squared Error:  0.6636468432894198\n",
      "Root Mean Squared Error:  0.8146452254137502\n",
      "R2 Score:  0.3279261868519444\n",
      "The accuracy of our model (based on R2 score) is 33.0%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.tree import DecisionTreeRegressor\n",
    "\n",
    "# Create and train the decision tree regressor \n",
    "dt_regressor = DecisionTreeRegressor(max_depth=5, random_state=42)\n",
    "dt_regressor.fit(X_train, y_train)\n",
    "\n",
    "# Make predictions with the trained model\n",
    "y_pred_dt = dt_regressor.predict(X_val)\n",
    "\n",
    "calculate_metrics(y_val, y_pred_dt, 0.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Random Forest Regressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Squared Error:  0.6560097700791785\n",
      "Root Mean Squared Error:  0.8099443006029355\n",
      "R2 Score:  0.33566023541346124\n",
      "The accuracy of our model (based on R2 score) is 34.0%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestRegressor\n",
    "\n",
    "# Create and train the Random Forest Regressor model\n",
    "rf_regressor = RandomForestRegressor(n_estimators=100, max_depth=5, random_state=42)\n",
    "rf_regressor.fit(X_train, y_train)\n",
    "\n",
    "# Make predictions with the trained model\n",
    "y_pred_rf = rf_regressor.predict(X_val)\n",
    "\n",
    "# Calculate the metrics\n",
    "calculate_metrics(y_val, y_pred_rf, 0.5)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Gradient Boosting Regressor "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Squared Error:  0.6322369524149647\n",
      "Root Mean Squared Error:  0.7951332922315382\n",
      "R2 Score:  0.3597349196802768\n",
      "The accuracy of our model (based on R2 score) is 36.0%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "\n",
    "# Create and train the gradient boosting model\n",
    "gb_regressor = GradientBoostingRegressor(n_estimators=100, learning_rate=0.1, max_depth=3, random_state=42)\n",
    "gb_regressor.fit(X_train, y_train)\n",
    "\n",
    "# Make predictions with the trained model\n",
    "y_pred_gb = gb_regressor.predict(X_val)\n",
    "\n",
    "# Calculate the metrics\n",
    "calculate_metrics(y_val, y_pred_gb, 0.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The models perform well, as the mean squared error is small compared to the size of input labels."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Model Selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Linear\n",
      "Mean Squared Error:  0.6537563034514501\n",
      "Root Mean Squared Error:  0.8085519794369748\n",
      "R2 Score:  0.34609805553203543\n",
      "The accuracy of our model (based on R2 score) is 35.0%\n",
      "\n",
      "\n",
      "SVR\n",
      "Mean Squared Error:  0.6357917338228121\n",
      "Root Mean Squared Error:  0.7973654957563765\n",
      "R2 Score:  0.36406662723020333\n",
      "The accuracy of our model (based on R2 score) is 36.0%\n",
      "\n",
      "\n",
      "Decision Tree\n",
      "Mean Squared Error:  0.6602257395621737\n",
      "Root Mean Squared Error:  0.812542761682223\n",
      "R2 Score:  0.3396271781881697\n",
      "The accuracy of our model (based on R2 score) is 34.0%\n",
      "\n",
      "\n",
      "Gredient Boosting\n",
      "\n",
      "Mean Squared Error:  0.6296948100657747\n",
      "Root Mean Squared Error:  0.7935331184429385\n",
      "R2 Score:  0.3701649092336835\n",
      "The accuracy of our model (based on R2 score) is 37.0%\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#test if overfitting using training data\n",
    "print('Linear')\n",
    "y_pred_linear_test = model.predict(X_train)\n",
    "calculate_metrics(y_train, y_pred_linear_test, 0.5)\n",
    "print('\\n')\n",
    "\n",
    "print('SVR')\n",
    "y_pred_svr_test = model_svr.predict(X_train)\n",
    "calculate_metrics(y_train, y_pred_svr_test, 0.5)\n",
    "print('\\n')\n",
    "\n",
    "print('Decision Tree')\n",
    "y_pred_dt_test = dt_regressor.predict(X_train)\n",
    "calculate_metrics(y_train, y_pred_dt_test, 0.5)\n",
    "print('\\n')\n",
    "\n",
    "print('Gredient Boosting\\n')\n",
    "y_pred_gb_test = gb_regressor.predict(X_train)\n",
    "calculate_metrics(y_train, y_pred_gb_test, 0.5)\n",
    "print('\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting the learning curves\n",
    "\n",
    "from sklearn.model_selection import LearningCurveDisplay, ShuffleSplit\n",
    "sns.set_style(\"whitegrid\")\n",
    "fig, ax = plt.subplots(nrows=1, ncols=3, figsize=(10, 6), sharey=True)\n",
    "\n",
    "common_params = {\n",
    "    \"X\": X_train_temp,\n",
    "    \"y\": y_train_temp,\n",
    "    \"train_sizes\": np.linspace(0.1, 1.0, 10),\n",
    "    \"cv\": ShuffleSplit(n_splits=5, test_size=0.2, random_state=0),\n",
    "    \"score_type\": \"both\",\n",
    "    \"n_jobs\": 4,\n",
    "    \"line_kw\": {\"marker\": \"o\"},\n",
    "    \"std_display_style\": \"fill_between\",\n",
    "    \"score_name\": \"Accuracy\",\n",
    "}\n",
    "\n",
    "for ax_idx, estimator in enumerate([model, model_svr, gb_regressor]):\n",
    "    LearningCurveDisplay.from_estimator(estimator, **common_params, ax=ax[ax_idx])\n",
    "    handles, label = ax[ax_idx].get_legend_handles_labels()\n",
    "    ax[ax_idx].legend(handles[:2], [\"Training Score\", \"Test Score\"])\n",
    "    ax[ax_idx].set_title(f\"Learning Curve for {estimator.__class__.__name__}\")\n",
    "    ax[ax_idx].set_ylim(0, 1)\n",
    "\n",
    "plt.tight_layout()\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Squared Error:  0.6549399069135492\n",
      "Root Mean Squared Error:  0.8092835763275745\n",
      "R2 Score:  0.33674368367294094\n",
      "The accuracy of our model (based on R2 score) is 34.0%\n"
     ]
    }
   ],
   "source": [
    "# Testing a linear SVR kernel\n",
    "\n",
    "model_svr = SVR(kernel = 'linear')\n",
    "model_svr.fit(X_train, y_train)\n",
    "\n",
    "y_pred_svr = model_svr.predict(X_val)\n",
    "\n",
    "calculate_metrics(y_val, y_pred_svr, 0.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Similar, just slightly worse performance than the default rbf kernel. This indicates that a linear relationship can be found among instances and targets."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cross-validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression\n",
      "count    10.000000\n",
      "mean      0.814865\n",
      "std       0.019014\n",
      "min       0.791322\n",
      "25%       0.801877\n",
      "50%       0.809136\n",
      "75%       0.830563\n",
      "max       0.848149\n",
      "dtype: float64\n",
      "\n",
      "\n",
      "Ridge\n",
      "count    10.000000\n",
      "mean      0.810783\n",
      "std       0.017807\n",
      "min       0.789923\n",
      "25%       0.800700\n",
      "50%       0.805556\n",
      "75%       0.810732\n",
      "max       0.847133\n",
      "dtype: float64\n",
      "\n",
      "\n",
      "GradientBoostingRegressor\n",
      "count    10.000000\n",
      "mean      0.804689\n",
      "std       0.016669\n",
      "min       0.785972\n",
      "25%       0.792295\n",
      "50%       0.801727\n",
      "75%       0.807054\n",
      "max       0.838721\n",
      "dtype: float64\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Running Cross Validation metrics\n",
    "from sklearn.model_selection import cross_val_score\n",
    "\n",
    "for index, estimator in enumerate([model, model_ridge, gb_regressor]):\n",
    "    mses = -cross_val_score(estimator, X_train, y_train, scoring=\"neg_root_mean_squared_error\", cv=10)\n",
    "    print(estimator.__class__.__name__)\n",
    "    print(pd.Series(mses).describe())\n",
    "    print('\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Model Hyperparameters: {'alpha': 10.0}\n",
      "Best Model Mean Squared Error on Test Set: 0.6476845409225035\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import Ridge\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "# Define the hyperparameters to tune on\n",
    "param_grid = {\n",
    "    'alpha': [0.001, 0.01, 0.1, 1.0, 10.0]\n",
    "}\n",
    "\n",
    "# Initialize the Ridge regressor\n",
    "ridge = Ridge()\n",
    "\n",
    "# Perform GridSearchCV for hyperparameter tuning\n",
    "grid_search = GridSearchCV(estimator=ridge, param_grid=param_grid, cv=5, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "grid_search.fit(X_train, y_train)\n",
    "\n",
    "# Get the best hyperparameters and the best estimator\n",
    "best_params = grid_search.best_params_\n",
    "best_estimator_linear = grid_search.best_estimator_\n",
    "\n",
    "# Evaluate the best model on the validation set\n",
    "y_pred = best_estimator_linear.predict(X_val)\n",
    "mse = mean_squared_error(y_val, y_pred)\n",
    "print(\"Best Model Hyperparameters:\", best_params)\n",
    "print(\"Best Model Mean Squared Error on Test Set:\", mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Model Hyperparameters: {'learning_rate': 0.1, 'max_depth': 4, 'min_samples_leaf': 2, 'min_samples_split': 2, 'n_estimators': 100}\n",
      "Best Model Mean Squared Error on Validation Set: 0.6253407555381476\n"
     ]
    }
   ],
   "source": [
    "# Define the hyperparameters to tune on\n",
    "param_grid = {\n",
    "    'n_estimators': [50, 100, 200],\n",
    "    'learning_rate': [0.01, 0.1, 0.5],\n",
    "    'max_depth': [3, 4, 5],\n",
    "    'min_samples_split': [2, 5, 10],\n",
    "    'min_samples_leaf': [1, 2, 4]\n",
    "}\n",
    "\n",
    "# Initialize the Gradient Boosting regressor with a Grid Search\n",
    "grid_search = GridSearchCV(estimator=gb_regressor, param_grid=param_grid, cv=5, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "grid_search.fit(X_train, y_train)\n",
    "\n",
    "best_params = grid_search.best_params_\n",
    "best_estimator_gb = grid_search.best_estimator_\n",
    "\n",
    "# Evaluate the best model on the validation set\n",
    "y_val_pred = best_estimator_gb.predict(X_val)\n",
    "val_mse = mean_squared_error(y_val, y_val_pred)\n",
    "print(\"Best Model Hyperparameters:\", best_params)\n",
    "print(\"Best Model Mean Squared Error on Validation Set:\", val_mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Squared Error:  0.6562231360214994\n",
      "Root Mean Squared Error:  0.810076006323789\n",
      "R2 Score:  0.35060100487941726\n",
      "The accuracy of our model (based on R2 score) is 35.0%\n",
      "Mean Squared Error:  0.6424088655997118\n",
      "Root Mean Squared Error:  0.8015041270010478\n",
      "R2 Score:  0.36427161909247496\n",
      "The accuracy of our model (based on R2 score) is 36.0%\n"
     ]
    }
   ],
   "source": [
    "# Evaluate the best model on the test set based on the best hyperparameters\n",
    "\n",
    "best_estimator_linear.fit(X_train, y_train)\n",
    "best_estimator_gb.fit(X_train, y_train)\n",
    "\n",
    "best_linear = best_estimator_linear.predict(X_test)\n",
    "best_gb = best_estimator_gb.predict(X_test)\n",
    "\n",
    "calculate_metrics(y_test, best_linear, 0.5)\n",
    "print('\\n')\n",
    "calculate_metrics(y_test, best_gb, 0.5)\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}