Content based Job recommender system.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import nltk\n",
    "from datetime import date\n",
    "from nltk.corpus import stopwords\n",
    "import re\n",
    "import string\n",
    "import time\n",
    "from nltk.stem import WordNetLemmatizer\n",
    "from nltk import word_tokenize\n",
    "from nltk.corpus import stopwords\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.cluster import KMeans\n",
    "from kmodes.kprototypes import KPrototypes\n",
    "\n",
    "# nltk.download('punkt')\n",
    "# nltk.download('stopwords')\n",
    "# nltk.download('wordnet')\n",
    "# nltk.download('averaged_perceptron_tagger')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Job_id</th>\n",
       "      <th>Job_Roles_</th>\n",
       "      <th>Country_Code_</th>\n",
       "      <th>glints_joborder_created_date</th>\n",
       "      <th>Skills_Required_</th>\n",
       "      <th>Year_of_Experience_Requirement__</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>Software_Engineer</td>\n",
       "      <td>MISSING_COUNTRY_CODE</td>\n",
       "      <td>2019-06-20T08:38:37.43</td>\n",
       "      <td>.NET, .NET, .NET_Core</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>Product_Manager</td>\n",
       "      <td>ID</td>\n",
       "      <td>2022-01-19T07:22:32.822</td>\n",
       "      <td>Jira, Scrum, Product_Management</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>Software_Engineer</td>\n",
       "      <td>ID</td>\n",
       "      <td>2021-10-22T09:48:58.893</td>\n",
       "      <td>React.js</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>Full-Stack_Developer</td>\n",
       "      <td>VN</td>\n",
       "      <td>2021-08-26T08:29:18.147</td>\n",
       "      <td>Vue.js, PHP</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>Research_&amp;_Development</td>\n",
       "      <td>ID</td>\n",
       "      <td>2021-09-15T06:43:24.815</td>\n",
       "      <td>Analysis, Analytical_Skills, Research</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Job_id              Job_Roles_         Country_Code_  \\\n",
       "0       1       Software_Engineer  MISSING_COUNTRY_CODE   \n",
       "1       2         Product_Manager                    ID   \n",
       "2       3       Software_Engineer                    ID   \n",
       "3       4    Full-Stack_Developer                    VN   \n",
       "4       5  Research_&_Development                    ID   \n",
       "\n",
       "  glints_joborder_created_date                       Skills_Required_  \\\n",
       "0       2019-06-20T08:38:37.43                  .NET, .NET, .NET_Core   \n",
       "1      2022-01-19T07:22:32.822        Jira, Scrum, Product_Management   \n",
       "2      2021-10-22T09:48:58.893                               React.js   \n",
       "3      2021-08-26T08:29:18.147                            Vue.js, PHP   \n",
       "4      2021-09-15T06:43:24.815  Analysis, Analytical_Skills, Research   \n",
       "\n",
       "   Year_of_Experience_Requirement__  \n",
       "0                                 6  \n",
       "1                                 3  \n",
       "2                                 2  \n",
       "3                                 2  \n",
       "4                                 1  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Jobs = pd.read_csv(\"Job_sample.csv\")\n",
    "Jobs.columns = [c.replace(' ', '_') for c in Jobs.columns]\n",
    "Jobs = Jobs.rename(columns={'id': 'Job_id'})\n",
    "Jobs.head()"
   ]
  },
  {
   "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>Candidate_id</th>\n",
       "      <th>glints_mergedprofile_country_code</th>\n",
       "      <th>Work_Experience_Title_</th>\n",
       "      <th>Skills_</th>\n",
       "      <th>Days_since_last_seen</th>\n",
       "      <th>Profile_Creation_Date</th>\n",
       "      <th>Days_of_experience</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>ID</td>\n",
       "      <td>Junior Aktivasi</td>\n",
       "      <td>Cisco_Packet_Tracer, Communication, CSS, Fiber...</td>\n",
       "      <td>323</td>\n",
       "      <td>2021-04-23T20:37:07.762</td>\n",
       "      <td>1081.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>SG</td>\n",
       "      <td>Human Resource intern Marketing intern</td>\n",
       "      <td>Administration, Communication, Detail_Oriented...</td>\n",
       "      <td>1</td>\n",
       "      <td>2022-01-28T09:40:38.837</td>\n",
       "      <td>166.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>ID</td>\n",
       "      <td>MISSING_WORK_EXPERIENCE_TITLES</td>\n",
       "      <td>Communication, English, Microsoft_Excel, Micro...</td>\n",
       "      <td>52</td>\n",
       "      <td>2022-03-18T06:15:53.818</td>\n",
       "      <td>-1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>ID</td>\n",
       "      <td>Kondektur Bus Sekolah</td>\n",
       "      <td>English, Presentation_Skills, Public_Speaking,...</td>\n",
       "      <td>816</td>\n",
       "      <td>2019-10-16T09:45:49.116</td>\n",
       "      <td>2068.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>ID</td>\n",
       "      <td>Bagian Umum Editor</td>\n",
       "      <td>Adobe_Photoshop, Corel_Draw_Suite, Editor, Med...</td>\n",
       "      <td>7</td>\n",
       "      <td>2022-05-18T10:42:22.182</td>\n",
       "      <td>365.25</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Candidate_id glints_mergedprofile_country_code  \\\n",
       "0             1                                ID   \n",
       "1             2                                SG   \n",
       "2             3                                ID   \n",
       "3             4                                ID   \n",
       "4             5                                ID   \n",
       "\n",
       "                   Work_Experience_Title_  \\\n",
       "0                         Junior Aktivasi   \n",
       "1  Human Resource intern Marketing intern   \n",
       "2          MISSING_WORK_EXPERIENCE_TITLES   \n",
       "3                   Kondektur Bus Sekolah   \n",
       "4                      Bagian Umum Editor   \n",
       "\n",
       "                                             Skills_  Days_since_last_seen  \\\n",
       "0  Cisco_Packet_Tracer, Communication, CSS, Fiber...                   323   \n",
       "1  Administration, Communication, Detail_Oriented...                     1   \n",
       "2  Communication, English, Microsoft_Excel, Micro...                    52   \n",
       "3  English, Presentation_Skills, Public_Speaking,...                   816   \n",
       "4  Adobe_Photoshop, Corel_Draw_Suite, Editor, Med...                     7   \n",
       "\n",
       "     Profile_Creation_Date  Days_of_experience  \n",
       "0  2021-04-23T20:37:07.762             1081.00  \n",
       "1  2022-01-28T09:40:38.837              166.00  \n",
       "2  2022-03-18T06:15:53.818               -1.00  \n",
       "3  2019-10-16T09:45:49.116             2068.75  \n",
       "4  2022-05-18T10:42:22.182              365.25  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Candidate = pd.read_csv(\"Candidate_sample.csv\")\n",
    "Candidate.columns = [c.replace(' ', '_') for c in Candidate.columns]\n",
    "Candidate = Candidate.rename(columns={'id': 'Candidate_id'})\n",
    "Candidate.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## EDA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<strong><span style=\"color:Red\"> Missing Values </span></strong>:  \n",
    "- Country codes for 58 of 161 Jobs are missing (36%)  \n",
    "- Candidate skills for 484 of 1960 candidates are missing (25%)\n",
    "- Candidate Work titles for 634 of 1960 candidates are missing  (32%)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Job_id</th>\n",
       "      <th>Job_Roles_</th>\n",
       "      <th>Country_Code_</th>\n",
       "      <th>glints_joborder_created_date</th>\n",
       "      <th>Skills_Required_</th>\n",
       "      <th>Year_of_Experience_Requirement__</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>Software_Engineer</td>\n",
       "      <td>MISSING_COUNTRY_CODE</td>\n",
       "      <td>2019-06-20T08:38:37.43</td>\n",
       "      <td>.NET, .NET, .NET_Core</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7</td>\n",
       "      <td>DevOps</td>\n",
       "      <td>MISSING_COUNTRY_CODE</td>\n",
       "      <td>2020-07-13T08:46:11.051</td>\n",
       "      <td>DevOps, Amazon_Web_Services_(AWS), Linux_Syste...</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10</td>\n",
       "      <td>Network_Engineer</td>\n",
       "      <td>MISSING_COUNTRY_CODE</td>\n",
       "      <td>2021-08-31T11:55:04.816</td>\n",
       "      <td>Network_Security, Switching, Routing, Networking</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>11</td>\n",
       "      <td>Android_Developer</td>\n",
       "      <td>MISSING_COUNTRY_CODE</td>\n",
       "      <td>2020-07-13T12:50:04.394</td>\n",
       "      <td>Kotlin</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>14</td>\n",
       "      <td>Digital_Marketing</td>\n",
       "      <td>MISSING_COUNTRY_CODE</td>\n",
       "      <td>2022-01-27T06:42:29.767</td>\n",
       "      <td>Management, Customer_Relationship_Management_(...</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Job_id         Job_Roles_         Country_Code_  \\\n",
       "0        1  Software_Engineer  MISSING_COUNTRY_CODE   \n",
       "6        7             DevOps  MISSING_COUNTRY_CODE   \n",
       "9       10   Network_Engineer  MISSING_COUNTRY_CODE   \n",
       "10      11  Android_Developer  MISSING_COUNTRY_CODE   \n",
       "13      14  Digital_Marketing  MISSING_COUNTRY_CODE   \n",
       "\n",
       "   glints_joborder_created_date  \\\n",
       "0        2019-06-20T08:38:37.43   \n",
       "6       2020-07-13T08:46:11.051   \n",
       "9       2021-08-31T11:55:04.816   \n",
       "10      2020-07-13T12:50:04.394   \n",
       "13      2022-01-27T06:42:29.767   \n",
       "\n",
       "                                     Skills_Required_  \\\n",
       "0                               .NET, .NET, .NET_Core   \n",
       "6   DevOps, Amazon_Web_Services_(AWS), Linux_Syste...   \n",
       "9    Network_Security, Switching, Routing, Networking   \n",
       "10                                             Kotlin   \n",
       "13  Management, Customer_Relationship_Management_(...   \n",
       "\n",
       "    Year_of_Experience_Requirement__  \n",
       "0                                  6  \n",
       "6                                  7  \n",
       "9                                  2  \n",
       "10                                 2  \n",
       "13                                 3  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Jobs.count()  #161\n",
    "Jobs[Jobs['Country_Code_'] == 'MISSING_COUNTRY_CODE'].count()   #58\n",
    "Jobs[Jobs['Country_Code_'] == 'MISSING_COUNTRY_CODE'].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>Candidate_id</th>\n",
       "      <th>glints_mergedprofile_country_code</th>\n",
       "      <th>Work_Experience_Title_</th>\n",
       "      <th>Skills_</th>\n",
       "      <th>Days_since_last_seen</th>\n",
       "      <th>Profile_Creation_Date</th>\n",
       "      <th>Days_of_experience</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>14</td>\n",
       "      <td>VN</td>\n",
       "      <td>Fundamental Business Analysis course Accor Pro...</td>\n",
       "      <td>MISSING_SKILLS</td>\n",
       "      <td>473</td>\n",
       "      <td>2021-02-28T05:06:12.657</td>\n",
       "      <td>1461.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>23</td>\n",
       "      <td>VN</td>\n",
       "      <td>MISSING_WORK_EXPERIENCE_TITLES</td>\n",
       "      <td>MISSING_SKILLS</td>\n",
       "      <td>123</td>\n",
       "      <td>2021-10-14T10:54:59.237</td>\n",
       "      <td>-1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>24</td>\n",
       "      <td>ID</td>\n",
       "      <td>MISSING_WORK_EXPERIENCE_TITLES</td>\n",
       "      <td>MISSING_SKILLS</td>\n",
       "      <td>180</td>\n",
       "      <td>2021-12-17T17:51:39.057</td>\n",
       "      <td>-1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>28</td>\n",
       "      <td>VN</td>\n",
       "      <td>MISSING_WORK_EXPERIENCE_TITLES</td>\n",
       "      <td>MISSING_SKILLS</td>\n",
       "      <td>136</td>\n",
       "      <td>2022-01-30T14:50:52.9</td>\n",
       "      <td>-1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>33</td>\n",
       "      <td>SG</td>\n",
       "      <td>MISSING_WORK_EXPERIENCE_TITLES</td>\n",
       "      <td>MISSING_SKILLS</td>\n",
       "      <td>1935</td>\n",
       "      <td>2017-02-26T11:17:09.266</td>\n",
       "      <td>-1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Candidate_id glints_mergedprofile_country_code  \\\n",
       "13            14                                VN   \n",
       "22            23                                VN   \n",
       "23            24                                ID   \n",
       "27            28                                VN   \n",
       "32            33                                SG   \n",
       "\n",
       "                               Work_Experience_Title_         Skills_  \\\n",
       "13  Fundamental Business Analysis course Accor Pro...  MISSING_SKILLS   \n",
       "22                     MISSING_WORK_EXPERIENCE_TITLES  MISSING_SKILLS   \n",
       "23                     MISSING_WORK_EXPERIENCE_TITLES  MISSING_SKILLS   \n",
       "27                     MISSING_WORK_EXPERIENCE_TITLES  MISSING_SKILLS   \n",
       "32                     MISSING_WORK_EXPERIENCE_TITLES  MISSING_SKILLS   \n",
       "\n",
       "    Days_since_last_seen    Profile_Creation_Date  Days_of_experience  \n",
       "13                   473  2021-02-28T05:06:12.657              1461.0  \n",
       "22                   123  2021-10-14T10:54:59.237                -1.0  \n",
       "23                   180  2021-12-17T17:51:39.057                -1.0  \n",
       "27                   136    2022-01-30T14:50:52.9                -1.0  \n",
       "32                  1935  2017-02-26T11:17:09.266                -1.0  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Candidate.count()  #1960\n",
    "Candidate[Candidate['Skills_'] == 'MISSING_SKILLS'].count()   #484\n",
    "Candidate[Candidate['Work_Experience_Title_'] == 'MISSING_WORK_EXPERIENCE_TITLES'].count()  #634\n",
    "Candidate[Candidate['Skills_'] == 'MISSING_SKILLS'].head() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Candidate_id</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Days_since_last_seen</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>-1</th>\n",
       "      <td>111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      Candidate_id\n",
       "Days_since_last_seen              \n",
       "-1                             111\n",
       " 0                              24\n",
       " 1                              22\n",
       " 2                              17\n",
       " 3                               7\n",
       " 4                               4\n",
       " 5                               9\n",
       " 6                              12\n",
       " 7                              10"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(Candidate.groupby(['Days_since_last_seen'])['Candidate_id'].count()).head(9)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature Engineering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<strong><span style=\"color:Blue\"> Time feature </span></strong>:  \n",
    "- Candidate experience in days is converted into years  \n",
    "- #Months when profile was created is obtained from Profile_Creation_Date\n",
    "- Days_since_last_seen is converted into Weeks_since_last_seen since to increase readability and if a candidate was last seen 2 days ago vs 6 days ago should not imply much difference"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>Job_id</th>\n",
       "      <th>Job_Roles_</th>\n",
       "      <th>Country_Code_</th>\n",
       "      <th>glints_joborder_created_date</th>\n",
       "      <th>Skills_Required_</th>\n",
       "      <th>Year_of_Experience_Requirement__</th>\n",
       "      <th>time</th>\n",
       "      <th>dates</th>\n",
       "      <th>Job_created_mo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>Software_Engineer</td>\n",
       "      <td>MISSING_COUNTRY_CODE</td>\n",
       "      <td>2019-06-20T08:38:37.43</td>\n",
       "      <td>.NET, .NET, .NET_Core</td>\n",
       "      <td>6</td>\n",
       "      <td>2019-06-20 08:38:37.430</td>\n",
       "      <td>2019-06-20</td>\n",
       "      <td>37</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>Product_Manager</td>\n",
       "      <td>ID</td>\n",
       "      <td>2022-01-19T07:22:32.822</td>\n",
       "      <td>Jira, Scrum, Product_Management</td>\n",
       "      <td>3</td>\n",
       "      <td>2022-01-19 07:22:32.822</td>\n",
       "      <td>2022-01-19</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>Software_Engineer</td>\n",
       "      <td>ID</td>\n",
       "      <td>2021-10-22T09:48:58.893</td>\n",
       "      <td>React.js</td>\n",
       "      <td>2</td>\n",
       "      <td>2021-10-22 09:48:58.893</td>\n",
       "      <td>2021-10-22</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>Full-Stack_Developer</td>\n",
       "      <td>VN</td>\n",
       "      <td>2021-08-26T08:29:18.147</td>\n",
       "      <td>Vue.js, PHP</td>\n",
       "      <td>2</td>\n",
       "      <td>2021-08-26 08:29:18.147</td>\n",
       "      <td>2021-08-26</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>Research_&amp;_Development</td>\n",
       "      <td>ID</td>\n",
       "      <td>2021-09-15T06:43:24.815</td>\n",
       "      <td>Analysis, Analytical_Skills, Research</td>\n",
       "      <td>1</td>\n",
       "      <td>2021-09-15 06:43:24.815</td>\n",
       "      <td>2021-09-15</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Job_id              Job_Roles_         Country_Code_  \\\n",
       "0       1       Software_Engineer  MISSING_COUNTRY_CODE   \n",
       "1       2         Product_Manager                    ID   \n",
       "2       3       Software_Engineer                    ID   \n",
       "3       4    Full-Stack_Developer                    VN   \n",
       "4       5  Research_&_Development                    ID   \n",
       "\n",
       "  glints_joborder_created_date                       Skills_Required_  \\\n",
       "0       2019-06-20T08:38:37.43                  .NET, .NET, .NET_Core   \n",
       "1      2022-01-19T07:22:32.822        Jira, Scrum, Product_Management   \n",
       "2      2021-10-22T09:48:58.893                               React.js   \n",
       "3      2021-08-26T08:29:18.147                            Vue.js, PHP   \n",
       "4      2021-09-15T06:43:24.815  Analysis, Analytical_Skills, Research   \n",
       "\n",
       "   Year_of_Experience_Requirement__                    time       dates  \\\n",
       "0                                 6 2019-06-20 08:38:37.430  2019-06-20   \n",
       "1                                 3 2022-01-19 07:22:32.822  2022-01-19   \n",
       "2                                 2 2021-10-22 09:48:58.893  2021-10-22   \n",
       "3                                 2 2021-08-26 08:29:18.147  2021-08-26   \n",
       "4                                 1 2021-09-15 06:43:24.815  2021-09-15   \n",
       "\n",
       "   Job_created_mo  \n",
       "0              37  \n",
       "1               6  \n",
       "2               9  \n",
       "3              11  \n",
       "4              10  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Jobs['time'] = pd.to_datetime(Jobs['glints_joborder_created_date'])\n",
    "Jobs['dates'] = Jobs['time'].dt.date\n",
    "Jobs['Job_created_mo'] = ((date.today()-Jobs['dates'])/np.timedelta64(1, 'M'))\n",
    "Jobs['Job_created_mo'] = Jobs['Job_created_mo'].astype(int)\n",
    "Jobs.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>Candidate_id</th>\n",
       "      <th>glints_mergedprofile_country_code</th>\n",
       "      <th>Work_Experience_Title_</th>\n",
       "      <th>Skills_</th>\n",
       "      <th>Days_since_last_seen</th>\n",
       "      <th>Profile_Creation_Date</th>\n",
       "      <th>Days_of_experience</th>\n",
       "      <th>time</th>\n",
       "      <th>dates</th>\n",
       "      <th>Profile_created_mo</th>\n",
       "      <th>Experience_in_yrs</th>\n",
       "      <th>Weeks_since_last_seen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>ID</td>\n",
       "      <td>Junior Aktivasi</td>\n",
       "      <td>Cisco_Packet_Tracer, Communication, CSS, Fiber...</td>\n",
       "      <td>323</td>\n",
       "      <td>2021-04-23T20:37:07.762</td>\n",
       "      <td>1081.00</td>\n",
       "      <td>2021-04-23 20:37:07.762</td>\n",
       "      <td>2021-04-23</td>\n",
       "      <td>15</td>\n",
       "      <td>2</td>\n",
       "      <td>46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>SG</td>\n",
       "      <td>Human Resource intern Marketing intern</td>\n",
       "      <td>Administration, Communication, Detail_Oriented...</td>\n",
       "      <td>1</td>\n",
       "      <td>2022-01-28T09:40:38.837</td>\n",
       "      <td>166.00</td>\n",
       "      <td>2022-01-28 09:40:38.837</td>\n",
       "      <td>2022-01-28</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>ID</td>\n",
       "      <td>MISSING_WORK_EXPERIENCE_TITLES</td>\n",
       "      <td>Communication, English, Microsoft_Excel, Micro...</td>\n",
       "      <td>52</td>\n",
       "      <td>2022-03-18T06:15:53.818</td>\n",
       "      <td>-1.00</td>\n",
       "      <td>2022-03-18 06:15:53.818</td>\n",
       "      <td>2022-03-18</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>ID</td>\n",
       "      <td>Kondektur Bus Sekolah</td>\n",
       "      <td>English, Presentation_Skills, Public_Speaking,...</td>\n",
       "      <td>816</td>\n",
       "      <td>2019-10-16T09:45:49.116</td>\n",
       "      <td>2068.75</td>\n",
       "      <td>2019-10-16 09:45:49.116</td>\n",
       "      <td>2019-10-16</td>\n",
       "      <td>33</td>\n",
       "      <td>5</td>\n",
       "      <td>116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>ID</td>\n",
       "      <td>Bagian Umum Editor</td>\n",
       "      <td>Adobe_Photoshop, Corel_Draw_Suite, Editor, Med...</td>\n",
       "      <td>7</td>\n",
       "      <td>2022-05-18T10:42:22.182</td>\n",
       "      <td>365.25</td>\n",
       "      <td>2022-05-18 10:42:22.182</td>\n",
       "      <td>2022-05-18</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Candidate_id glints_mergedprofile_country_code  \\\n",
       "0             1                                ID   \n",
       "1             2                                SG   \n",
       "2             3                                ID   \n",
       "3             4                                ID   \n",
       "4             5                                ID   \n",
       "\n",
       "                   Work_Experience_Title_  \\\n",
       "0                         Junior Aktivasi   \n",
       "1  Human Resource intern Marketing intern   \n",
       "2          MISSING_WORK_EXPERIENCE_TITLES   \n",
       "3                   Kondektur Bus Sekolah   \n",
       "4                      Bagian Umum Editor   \n",
       "\n",
       "                                             Skills_  Days_since_last_seen  \\\n",
       "0  Cisco_Packet_Tracer, Communication, CSS, Fiber...                   323   \n",
       "1  Administration, Communication, Detail_Oriented...                     1   \n",
       "2  Communication, English, Microsoft_Excel, Micro...                    52   \n",
       "3  English, Presentation_Skills, Public_Speaking,...                   816   \n",
       "4  Adobe_Photoshop, Corel_Draw_Suite, Editor, Med...                     7   \n",
       "\n",
       "     Profile_Creation_Date  Days_of_experience                    time  \\\n",
       "0  2021-04-23T20:37:07.762             1081.00 2021-04-23 20:37:07.762   \n",
       "1  2022-01-28T09:40:38.837              166.00 2022-01-28 09:40:38.837   \n",
       "2  2022-03-18T06:15:53.818               -1.00 2022-03-18 06:15:53.818   \n",
       "3  2019-10-16T09:45:49.116             2068.75 2019-10-16 09:45:49.116   \n",
       "4  2022-05-18T10:42:22.182              365.25 2022-05-18 10:42:22.182   \n",
       "\n",
       "        dates  Profile_created_mo  Experience_in_yrs  Weeks_since_last_seen  \n",
       "0  2021-04-23                  15                  2                     46  \n",
       "1  2022-01-28                   6                  0                      0  \n",
       "2  2022-03-18                   4                  0                      7  \n",
       "3  2019-10-16                  33                  5                    116  \n",
       "4  2022-05-18                   2                  1                      1  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Candidate['time'] = pd.to_datetime(Candidate['Profile_Creation_Date'])\n",
    "Candidate['dates'] = Candidate['time'].dt.date\n",
    "Candidate['Profile_created_mo'] = ((date.today()-Candidate['dates'])/np.timedelta64(1, 'M'))\n",
    "Candidate['Profile_created_mo'] = Candidate['Profile_created_mo'].astype(int)\n",
    "Candidate['Experience_in_yrs'] = (Candidate['Days_of_experience']/365).astype(int)\n",
    "Candidate['Weeks_since_last_seen'] = (Candidate['Days_since_last_seen']/7).astype(int)\n",
    "Candidate.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 1: Methodology chosen - Content based Recommender System¶"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-info\">\n",
    "<b> Why Content Based System </b></div>\n",
    "\n",
    "- There are broadly 3 types of recommender systems - Collaborative Filtering , Content based and Hybrid System.\n",
    "- For Collaborative filtering & Hybrid System , an implicit feedback is required which is often a numerical value which denotes affinity of a user (here Candidate) to an item (here Job)\n",
    "- In the shared dataset, we do have information like profile_created , Last_seen and Candidate Experiences alone but that cannot help us match required Job with the suitable Candidate as per their Skills & Job title \n",
    "- Hence it makes sense to use Content based RecSys since it can help us match the Job with a suitable candidate basis her Skills and other text information"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-info\">\n",
    "<b> How to implement </b></div>\n",
    "\n",
    "- Get relevant text data for both Jobs (Job_role , Skills required , Country code) and Candidates (Work title , Skills , Country code) dataset\n",
    "- Get their feature vector using any NLP techniques like BOW , TF-IDF or Word2vec.\n",
    "- Calculate cosine similiary between each Job vector vs Candidate vector and sort it basis these scores\n",
    "- Filter the top-N Candidates bases their score\n",
    "- As an additional filter , we can also filter candidates basis their matching country code with the Job country code and last_seen_activity"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create corpus"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>Job_id</th>\n",
       "      <th>Country_Code_</th>\n",
       "      <th>Job_Roles_</th>\n",
       "      <th>Jtext</th>\n",
       "      <th>Job_created_mo</th>\n",
       "      <th>Skills_Required_</th>\n",
       "      <th>Year_of_Experience_Requirement__</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>MISSING_COUNTRY_CODE</td>\n",
       "      <td>Software Engineer</td>\n",
       "      <td>Software Engineer .NET, .NET, .NET_Core MISSIN...</td>\n",
       "      <td>37</td>\n",
       "      <td>.NET, .NET, .NET_Core</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>ID</td>\n",
       "      <td>Product Manager</td>\n",
       "      <td>Product Manager Jira, Scrum, Product_Managemen...</td>\n",
       "      <td>6</td>\n",
       "      <td>Jira, Scrum, Product_Management</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>ID</td>\n",
       "      <td>Software Engineer</td>\n",
       "      <td>Software Engineer React.js ID</td>\n",
       "      <td>9</td>\n",
       "      <td>React.js</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>VN</td>\n",
       "      <td>Full Stack Developer</td>\n",
       "      <td>Full Stack Developer Vue.js, PHP VN</td>\n",
       "      <td>11</td>\n",
       "      <td>Vue.js, PHP</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>ID</td>\n",
       "      <td>Research &amp; Development</td>\n",
       "      <td>Research &amp; Development Analysis, Analytical_Sk...</td>\n",
       "      <td>10</td>\n",
       "      <td>Analysis, Analytical_Skills, Research</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Job_id         Country_Code_              Job_Roles_  \\\n",
       "0       1  MISSING_COUNTRY_CODE       Software Engineer   \n",
       "1       2                    ID         Product Manager   \n",
       "2       3                    ID       Software Engineer   \n",
       "3       4                    VN    Full Stack Developer   \n",
       "4       5                    ID  Research & Development   \n",
       "\n",
       "                                               Jtext  Job_created_mo  \\\n",
       "0  Software Engineer .NET, .NET, .NET_Core MISSIN...              37   \n",
       "1  Product Manager Jira, Scrum, Product_Managemen...               6   \n",
       "2                      Software Engineer React.js ID               9   \n",
       "3                Full Stack Developer Vue.js, PHP VN              11   \n",
       "4  Research & Development Analysis, Analytical_Sk...              10   \n",
       "\n",
       "                        Skills_Required_  Year_of_Experience_Requirement__  \n",
       "0                  .NET, .NET, .NET_Core                                 6  \n",
       "1        Jira, Scrum, Product_Management                                 3  \n",
       "2                               React.js                                 2  \n",
       "3                            Vue.js, PHP                                 2  \n",
       "4  Analysis, Analytical_Skills, Research                                 1  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Jobs['Job_Roles_'] = Jobs['Job_Roles_'].replace('_',' ', regex=True)\n",
    "Jobs['Job_Roles_'] = Jobs['Job_Roles_'].replace('-',' ', regex=True)\n",
    "\n",
    "Jobs[\"Jtext\"] = Jobs[\"Job_Roles_\"].map(str) + \" \" + Jobs[\"Skills_Required_\"] +\" \"+ Jobs[\"Country_Code_\"]\n",
    "Jobs = Jobs[['Job_id','Country_Code_','Job_Roles_','Jtext','Job_created_mo','Skills_Required_','Year_of_Experience_Requirement__']]\n",
    "Jobs.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>Candidate_id</th>\n",
       "      <th>glints_mergedprofile_country_code</th>\n",
       "      <th>Skills_</th>\n",
       "      <th>Work_Experience_Title_</th>\n",
       "      <th>Ctext</th>\n",
       "      <th>Profile_created_mo</th>\n",
       "      <th>Experience_in_yrs</th>\n",
       "      <th>Weeks_since_last_seen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>ID</td>\n",
       "      <td>Cisco_Packet_Tracer, Communication, CSS, Fiber...</td>\n",
       "      <td>Junior Aktivasi</td>\n",
       "      <td>Junior Aktivasi Cisco_Packet_Tracer, Communica...</td>\n",
       "      <td>15</td>\n",
       "      <td>2</td>\n",
       "      <td>46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>SG</td>\n",
       "      <td>Administration, Communication, Detail_Oriented...</td>\n",
       "      <td>Human Resource intern Marketing intern</td>\n",
       "      <td>Human Resource intern Marketing intern Adminis...</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>ID</td>\n",
       "      <td>Communication, English, Microsoft_Excel, Micro...</td>\n",
       "      <td>MISSING_WORK_EXPERIENCE_TITLES</td>\n",
       "      <td>MISSING_WORK_EXPERIENCE_TITLES Communication, ...</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>ID</td>\n",
       "      <td>English, Presentation_Skills, Public_Speaking,...</td>\n",
       "      <td>Kondektur Bus Sekolah</td>\n",
       "      <td>Kondektur Bus Sekolah English, Presentation_Sk...</td>\n",
       "      <td>33</td>\n",
       "      <td>5</td>\n",
       "      <td>116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>ID</td>\n",
       "      <td>Adobe_Photoshop, Corel_Draw_Suite, Editor, Med...</td>\n",
       "      <td>Bagian Umum Editor</td>\n",
       "      <td>Bagian Umum Editor Adobe_Photoshop, Corel_Draw...</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Candidate_id glints_mergedprofile_country_code  \\\n",
       "0             1                                ID   \n",
       "1             2                                SG   \n",
       "2             3                                ID   \n",
       "3             4                                ID   \n",
       "4             5                                ID   \n",
       "\n",
       "                                             Skills_  \\\n",
       "0  Cisco_Packet_Tracer, Communication, CSS, Fiber...   \n",
       "1  Administration, Communication, Detail_Oriented...   \n",
       "2  Communication, English, Microsoft_Excel, Micro...   \n",
       "3  English, Presentation_Skills, Public_Speaking,...   \n",
       "4  Adobe_Photoshop, Corel_Draw_Suite, Editor, Med...   \n",
       "\n",
       "                   Work_Experience_Title_  \\\n",
       "0                         Junior Aktivasi   \n",
       "1  Human Resource intern Marketing intern   \n",
       "2          MISSING_WORK_EXPERIENCE_TITLES   \n",
       "3                   Kondektur Bus Sekolah   \n",
       "4                      Bagian Umum Editor   \n",
       "\n",
       "                                               Ctext  Profile_created_mo  \\\n",
       "0  Junior Aktivasi Cisco_Packet_Tracer, Communica...                  15   \n",
       "1  Human Resource intern Marketing intern Adminis...                   6   \n",
       "2  MISSING_WORK_EXPERIENCE_TITLES Communication, ...                   4   \n",
       "3  Kondektur Bus Sekolah English, Presentation_Sk...                  33   \n",
       "4  Bagian Umum Editor Adobe_Photoshop, Corel_Draw...                   2   \n",
       "\n",
       "   Experience_in_yrs  Weeks_since_last_seen  \n",
       "0                  2                     46  \n",
       "1                  0                      0  \n",
       "2                  0                      7  \n",
       "3                  5                    116  \n",
       "4                  1                      1  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Candidate[\"Ctext\"] = Candidate[\"Work_Experience_Title_\"].map(str) + \" \" + Candidate[\"Skills_\"] +\" \"+ Candidate[\"glints_mergedprofile_country_code\"]\n",
    "Candidate = Candidate[['Candidate_id','glints_mergedprofile_country_code','Skills_','Work_Experience_Title_','Ctext',\n",
    "                       'Profile_created_mo','Experience_in_yrs','Weeks_since_last_seen']]\n",
    "\n",
    "Candidate.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Clean corpus"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>Job_id</th>\n",
       "      <th>Country_Code_</th>\n",
       "      <th>Job_Roles_</th>\n",
       "      <th>Jtext</th>\n",
       "      <th>Job_created_mo</th>\n",
       "      <th>Skills_Required_</th>\n",
       "      <th>Year_of_Experience_Requirement__</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>MISSING_COUNTRY_CODE</td>\n",
       "      <td>Software Engineer</td>\n",
       "      <td>software engineer net net net_core missing_cou...</td>\n",
       "      <td>37</td>\n",
       "      <td>.NET, .NET, .NET_Core</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>ID</td>\n",
       "      <td>Product Manager</td>\n",
       "      <td>product manager jira scrum product_management</td>\n",
       "      <td>6</td>\n",
       "      <td>Jira, Scrum, Product_Management</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>ID</td>\n",
       "      <td>Software Engineer</td>\n",
       "      <td>software engineer react</td>\n",
       "      <td>9</td>\n",
       "      <td>React.js</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>VN</td>\n",
       "      <td>Full Stack Developer</td>\n",
       "      <td>full stack developer vue php</td>\n",
       "      <td>11</td>\n",
       "      <td>Vue.js, PHP</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>ID</td>\n",
       "      <td>Research &amp; Development</td>\n",
       "      <td>research development analysis analytical_skill...</td>\n",
       "      <td>10</td>\n",
       "      <td>Analysis, Analytical_Skills, Research</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Job_id         Country_Code_              Job_Roles_  \\\n",
       "0       1  MISSING_COUNTRY_CODE       Software Engineer   \n",
       "1       2                    ID         Product Manager   \n",
       "2       3                    ID       Software Engineer   \n",
       "3       4                    VN    Full Stack Developer   \n",
       "4       5                    ID  Research & Development   \n",
       "\n",
       "                                               Jtext  Job_created_mo  \\\n",
       "0  software engineer net net net_core missing_cou...              37   \n",
       "1      product manager jira scrum product_management               6   \n",
       "2                            software engineer react               9   \n",
       "3                       full stack developer vue php              11   \n",
       "4  research development analysis analytical_skill...              10   \n",
       "\n",
       "                        Skills_Required_  Year_of_Experience_Requirement__  \n",
       "0                  .NET, .NET, .NET_Core                                 6  \n",
       "1        Jira, Scrum, Product_Management                                 3  \n",
       "2                               React.js                                 2  \n",
       "3                            Vue.js, PHP                                 2  \n",
       "4  Analysis, Analytical_Skills, Research                                 1  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stop = stopwords.words('english')\n",
    "stop_words_ = set(stopwords.words('english'))\n",
    "wn = WordNetLemmatizer()\n",
    "\n",
    "def black_txt(token):\n",
    "    return  token not in stop_words_ and token not in list(string.punctuation)  and len(token)>2   \n",
    "  \n",
    "def clean_txt(text):\n",
    "  clean_text = []\n",
    "  clean_text2 = []\n",
    "  text = re.sub(\"'\", \"\",text)\n",
    "  text=re.sub(\"(\\\\d|\\\\W)+\",\" \",text) \n",
    "  text = text.replace(\"nbsp\", \"\")\n",
    "  clean_text = [ wn.lemmatize(word, pos=\"v\") for word in word_tokenize(text.lower()) if black_txt(word)]\n",
    "  clean_text2 = [word for word in clean_text if black_txt(word)]\n",
    "  return \" \".join(clean_text2)\n",
    "\n",
    "Jobs['Jtext'] = Jobs['Jtext'].apply(clean_txt)\n",
    "Jobs.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>Candidate_id</th>\n",
       "      <th>glints_mergedprofile_country_code</th>\n",
       "      <th>Skills_</th>\n",
       "      <th>Work_Experience_Title_</th>\n",
       "      <th>Ctext</th>\n",
       "      <th>Profile_created_mo</th>\n",
       "      <th>Experience_in_yrs</th>\n",
       "      <th>Weeks_since_last_seen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>291</th>\n",
       "      <td>292</td>\n",
       "      <td>ID</td>\n",
       "      <td>Communication, Marketing, Microsoft_Office, My...</td>\n",
       "      <td>Software Development Engineer Business Develop...</td>\n",
       "      <td>software development engineer business develop...</td>\n",
       "      <td>30</td>\n",
       "      <td>6</td>\n",
       "      <td>119</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Candidate_id glints_mergedprofile_country_code  \\\n",
       "291           292                                ID   \n",
       "\n",
       "                                               Skills_  \\\n",
       "291  Communication, Marketing, Microsoft_Office, My...   \n",
       "\n",
       "                                Work_Experience_Title_  \\\n",
       "291  Software Development Engineer Business Develop...   \n",
       "\n",
       "                                                 Ctext  Profile_created_mo  \\\n",
       "291  software development engineer business develop...                  30   \n",
       "\n",
       "     Experience_in_yrs  Weeks_since_last_seen  \n",
       "291                  6                    119  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stop = stopwords.words('english')\n",
    "stop_words_ = set(stopwords.words('english'))\n",
    "wn = WordNetLemmatizer()\n",
    "\n",
    "def black_txt(token):\n",
    "    return  token not in stop_words_ and token not in list(string.punctuation)  and len(token)>2   \n",
    "  \n",
    "def clean_txt(text):\n",
    "  clean_text = []\n",
    "  clean_text2 = []\n",
    "  text = re.sub(\"'\", \"\",text)\n",
    "  text=re.sub(\"(\\\\d|\\\\W)+\",\" \",text) \n",
    "  text = text.replace(\"nbsp\", \"\")\n",
    "  clean_text = [ wn.lemmatize(word, pos=\"v\") for word in word_tokenize(text.lower()) if black_txt(word)]\n",
    "  clean_text2 = [word for word in clean_text if black_txt(word)]\n",
    "  return \" \".join(clean_text2)\n",
    "\n",
    "Candidate['Ctext'] = Candidate['Ctext'].apply(clean_txt)\n",
    "Candidate[Candidate['Candidate_id']==292].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TF IDF vectors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "tfidf_vectorizer = TfidfVectorizer()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1960, 4162)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tfidf_candidate = tfidf_vectorizer.fit_transform((Candidate['Ctext'])) #fitting and transforming the vector\n",
    "tfidf_candidate.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 2 : Handling missing values¶"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-info\">\n",
    "\n",
    "- Calculate Cosine similarity between different Candidate TF-IDF vectors\n",
    "- Pick the most similar candidate as per these scores, chances are the candidates with the missing values will have the same Work experince, Skills etc as the most similar Candidate vector  </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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_id</th>\n",
       "      <th>Matched_id</th>\n",
       "      <th>Work_Experience_Title_</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12</td>\n",
       "      <td>118</td>\n",
       "      <td>Assistant for Accounting Supervisor</td>\n",
       "      <td>0.641984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>372</th>\n",
       "      <td>19</td>\n",
       "      <td>295</td>\n",
       "      <td>Marketing</td>\n",
       "      <td>0.380932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>20</td>\n",
       "      <td>961</td>\n",
       "      <td>Staff Accounting</td>\n",
       "      <td>0.769941</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>304</th>\n",
       "      <td>23</td>\n",
       "      <td>778</td>\n",
       "      <td>R&amp;D</td>\n",
       "      <td>0.747895</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Missing_id  Matched_id               Work_Experience_Title_     score\n",
       "2            12         118  Assistant for Accounting Supervisor  0.641984\n",
       "372          19         295                            Marketing  0.380932\n",
       "4            20         961                     Staff Accounting  0.769941\n",
       "304          23         778                                  R&D  0.747895"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def Missing_values_imputer(df,Id_col,Missing_column,Missing_column_values,tf_idf,text):\n",
    "    \n",
    "    def Nearest_value_finder(top, df, scores,Missing_Id):\n",
    "        nearest = pd.DataFrame(columns = ['Missing_id','Matched_id',Missing_column,'score'])\n",
    "        count = 0\n",
    "        for i in top:\n",
    "            nearest.at[count, 'Missing_id'] = Missing_Id\n",
    "            nearest.at[count, 'Matched_id'] = df[Id_col][i]\n",
    "            nearest.at[count, Missing_column] = df[Missing_column][i]\n",
    "            nearest.at[count, 'score'] =  scores[count]\n",
    "            count += 1\n",
    "        return nearest\n",
    "\n",
    "    nearest_df = pd.DataFrame(['Missing_id','Matched_id',Missing_column,'score'])\n",
    "    df_list = df[df[Missing_column] == Missing_column_values]\n",
    "    df_list = df_list[Id_col][1:5].unique()\n",
    "    \n",
    "#     return df_list\n",
    "\n",
    "    for ID in df_list:\n",
    "        missing_tf_idf = tfidf_vectorizer.transform(df[df[Id_col] == ID][text])\n",
    "        cos_similarity_tfidf = map(lambda x: cosine_similarity(missing_tf_idf, x),tf_idf)\n",
    "        output2 = list(cos_similarity_tfidf)\n",
    "\n",
    "        top = sorted(range(len(output2)), key=lambda i: output2[i], reverse=True)\n",
    "        list_scores = [output2[i][0][0] for i in top]\n",
    "        nearest_df = nearest_df.append(pd.DataFrame(Nearest_value_finder(top,df, list_scores,ID)))\n",
    "\n",
    "    nearest_df = nearest_df[['Missing_id','Matched_id',Missing_column,'score']]\n",
    "    nearest_df = nearest_df.dropna()\n",
    "    nearest_df.fillna(0,inplace=True)\n",
    "    nearest_df = nearest_df[nearest_df[Missing_column] != Missing_column_values]\n",
    "    nearest_df = nearest_df.groupby(['Missing_id']).head(1)\n",
    "    return nearest_df\n",
    "\n",
    "Missing_WorkEx = Missing_values_imputer(Candidate,'Candidate_id','Work_Experience_Title_','MISSING_WORK_EXPERIENCE_TITLES',tfidf_candidate,'Ctext')\n",
    "# Missing_Candidate_skills = Missing_values_imputer(Candidate,'Candidate_id','Skills_','MISSING_SKILLS',tfidf_candidate,'Ctext')\n",
    "# Missing_job_skills = Missing_values_imputer(Jobs,'Job_id','Skills_Required_','MISSING_SKILLS',tfidf_job,'Jtext')\n",
    "# Missing_country = Missing_values_imputer(Jobs,'Job_id','Country_code_','MISSING_COUNTRY_CODE',tfidf_job,'Jtext')\n",
    "\n",
    "Missing_WorkEx.head(10)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Here Candidate_id with ID 12 has missing Work_exp but it should be most similar to Candidate_id 118 as shown below"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "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>Candidate_id</th>\n",
       "      <th>glints_mergedprofile_country_code</th>\n",
       "      <th>Skills_</th>\n",
       "      <th>Work_Experience_Title_</th>\n",
       "      <th>Ctext</th>\n",
       "      <th>Profile_created_mo</th>\n",
       "      <th>Experience_in_yrs</th>\n",
       "      <th>Weeks_since_last_seen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>12</td>\n",
       "      <td>ID</td>\n",
       "      <td>Accounting, Communication, Detail_Oriented, Fi...</td>\n",
       "      <td>MISSING_WORK_EXPERIENCE_TITLES</td>\n",
       "      <td>missing_work_experience_titles account communi...</td>\n",
       "      <td>33</td>\n",
       "      <td>0</td>\n",
       "      <td>136</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Candidate_id glints_mergedprofile_country_code  \\\n",
       "11            12                                ID   \n",
       "\n",
       "                                              Skills_  \\\n",
       "11  Accounting, Communication, Detail_Oriented, Fi...   \n",
       "\n",
       "            Work_Experience_Title_  \\\n",
       "11  MISSING_WORK_EXPERIENCE_TITLES   \n",
       "\n",
       "                                                Ctext  Profile_created_mo  \\\n",
       "11  missing_work_experience_titles account communi...                  33   \n",
       "\n",
       "    Experience_in_yrs  Weeks_since_last_seen  \n",
       "11                  0                    136  "
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Candidate[Candidate['Candidate_id'] == 12].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "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>Candidate_id</th>\n",
       "      <th>glints_mergedprofile_country_code</th>\n",
       "      <th>Skills_</th>\n",
       "      <th>Work_Experience_Title_</th>\n",
       "      <th>Ctext</th>\n",
       "      <th>Profile_created_mo</th>\n",
       "      <th>Experience_in_yrs</th>\n",
       "      <th>Weeks_since_last_seen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>117</th>\n",
       "      <td>118</td>\n",
       "      <td>ID</td>\n",
       "      <td>Accounting, Communication, Detail_Oriented, Fi...</td>\n",
       "      <td>Assistant for Accounting Supervisor</td>\n",
       "      <td>assistant account supervisor account communica...</td>\n",
       "      <td>17</td>\n",
       "      <td>8</td>\n",
       "      <td>68</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Candidate_id glints_mergedprofile_country_code  \\\n",
       "117           118                                ID   \n",
       "\n",
       "                                               Skills_  \\\n",
       "117  Accounting, Communication, Detail_Oriented, Fi...   \n",
       "\n",
       "                  Work_Experience_Title_  \\\n",
       "117  Assistant for Accounting Supervisor   \n",
       "\n",
       "                                                 Ctext  Profile_created_mo  \\\n",
       "117  assistant account supervisor account communica...                  17   \n",
       "\n",
       "     Experience_in_yrs  Weeks_since_last_seen  \n",
       "117                  8                     68  "
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Candidate[Candidate['Candidate_id'] == 118].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Top N recommended candidates for each Job using cosine similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_recommendation(top, df_all, scores,job_id):\n",
    "  recommendation = pd.DataFrame(columns = ['Candidate_id', 'Job_id',  'Job_Roles_', 'score'])\n",
    "  count = 0\n",
    "  for i in top:\n",
    "      recommendation.at[count, 'Job_id'] = job_id\n",
    "      recommendation.at[count, 'Candidate_id'] = df_all['Candidate_id'][i]\n",
    "      #recommendation.at[count, 'Job_Roles_'] = df_all['Job_Roles_'][i]\n",
    "      recommendation.at[count, 'score'] =  scores[count]\n",
    "      count += 1\n",
    "  return recommendation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>Job_id</th>\n",
       "      <th>Candidate_id</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>979</td>\n",
       "      <td>0.606158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>71</td>\n",
       "      <td>0.492805</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>293</td>\n",
       "      <td>0.373916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1896</td>\n",
       "      <td>0.364500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>898</td>\n",
       "      <td>0.349587</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Job_id  Candidate_id     score\n",
       "0       1           979  0.606158\n",
       "1       1            71  0.492805\n",
       "2       1           293  0.373916\n",
       "3       1          1896  0.364500\n",
       "4       1           898  0.349587"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rec_df = pd.DataFrame(['Candidate_id','Job_id','Job_Roles_','score'])\n",
    "\n",
    "job_list = Jobs['Job_id'].unique()\n",
    "#job_list = Jobs['Job_id'][1:10].unique()\n",
    "\n",
    "for job_id in job_list:\n",
    "    tfidf_job = tfidf_vectorizer.transform(Jobs[Jobs['Job_id'] == job_id]['Jtext'])\n",
    "    cos_similarity_tfidf = map(lambda x: cosine_similarity(tfidf_job, x),tfidf_candidate)\n",
    "    output2 = list(cos_similarity_tfidf)\n",
    "\n",
    "    top = sorted(range(len(output2)), key=lambda i: output2[i], reverse=True)\n",
    "    list_scores = [output2[i][0][0] for i in top]\n",
    "    rec_df = rec_df.append(pd.DataFrame(get_recommendation(top,Candidate, list_scores,job_id)))\n",
    "\n",
    "rec_df = rec_df[['Job_id','Candidate_id','score']]\n",
    "rec_df = rec_df.dropna()\n",
    "rec_df.fillna(0,inplace=True)\n",
    "#rec_df = rec_df[rec_df['score'] > 0]\n",
    "rec_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Append last seen, profile created , experience and job posted timetamps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "Final_df = Jobs.merge(rec_df,how =\"left\",on='Job_id')\n",
    "Final_df = Final_df.merge(Candidate,how =\"left\",on='Candidate_id')\n",
    "Final_df = Final_df[['Job_id','Candidate_id','Job_Roles_','Country_Code_','glints_mergedprofile_country_code'\n",
    "                    ,'Work_Experience_Title_','Year_of_Experience_Requirement__','Experience_in_yrs',\n",
    "                     'Skills_Required_','Skills_','Profile_created_mo','Weeks_since_last_seen','score']]\n",
    "Final_df = Final_df.rename(columns={'Job_Roles_': 'Job_Role',\n",
    "                                    'Country_Code_' : 'Job_country',\n",
    "                                   'glints_mergedprofile_country_code' : 'Candidate_country' ,\n",
    "                                   'Work_Experience_Title_' : 'Candidate_job_title',\n",
    "                                   'Year_of_Experience_Requirement__' : 'Exp_required',\n",
    "                                   'Skills_Required_' : 'Skills_required',\n",
    "                                   'Skills_' : 'Candidate_skills',\n",
    "                                   'Experience_in_yrs' : 'Candidate_exp'})\n",
    "Final_df.head()\n",
    "Final_df.to_csv('Recommendations.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Final_df['Candidate_id'].isna().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 3 : Case 1 (Predict Likelihood of Candidates being open to ANY Job postings using Kmeans)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-info\">\n",
    "<b>Making predictions basis just the Last_seen information since it would denote how proactively the candidate is looking for a job change </b> </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-21-b271396939fe>:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  Candidate_last_seen.drop_duplicates(inplace=True)\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>Weeks_since_last_seen</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Candidate_id</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Weeks_since_last_seen\n",
       "Candidate_id                       \n",
       "1                                46\n",
       "2                                 0\n",
       "3                                 7\n",
       "4                               116\n",
       "5                                 1"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Candidate_last_seen = Candidate[['Candidate_id','Weeks_since_last_seen']]\n",
    "Candidate_last_seen.drop_duplicates(inplace=True)\n",
    "Candidate_last_seen.set_index('Candidate_id',inplace=True)\n",
    "Candidate_last_seen.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## K-means for obtaining clusters"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Outlier Treatment"
   ]
  },
  {
   "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>Weeks_since_last_seen</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Candidate_id</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.213412</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.923945</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.815821</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.867834</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>-0.908499</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Weeks_since_last_seen\n",
       "Candidate_id                       \n",
       "1                         -0.213412\n",
       "2                         -0.923945\n",
       "3                         -0.815821\n",
       "4                          0.867834\n",
       "5                         -0.908499"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = Candidate_last_seen[['Weeks_since_last_seen']]\n",
    "\n",
    "#Missing values\n",
    "df = df.fillna(0)\n",
    "df = df.replace([np.inf, -np.inf], np.nan)\n",
    "df.isnull().values.any()\n",
    "\n",
    "#standardize\n",
    "#df_std = pd.DataFrame(StandardScaler().fit_transform(df))\n",
    "\n",
    "df_std = pd.DataFrame(StandardScaler().fit_transform(df), index=df.index, columns=df.columns)\n",
    "\n",
    "#Treat outliers (Removing obs that fall beyond 3 stand dev)\n",
    "for col in df_std.columns:\n",
    "    upper = 3\n",
    "    lower = -3\n",
    "    df_std[col][df_std[col] < lower] = lower\n",
    "    df_std[col][df_std[col] > upper] = upper\n",
    "\n",
    "df_std.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Elbow method for clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df_std[['Weeks_since_last_seen']].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "wcss = []\n",
    "for i in range(4, 10):\n",
    "    kmeans = KMeans(n_clusters=i, init='k-means++', max_iter=300, random_state=0)\n",
    "    kmeans.fit(X)\n",
    "    wcss.append(kmeans.inertia_)\n",
    "plt.plot(range(4,10), wcss)\n",
    "plt.title('Elbow Method')\n",
    "plt.xlabel('Number of clusters')\n",
    "plt.ylabel('WCSS')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2, 0, 0, ..., 3, 3, 3], dtype=int32)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = KMeans(n_clusters=6,init='k-means++', max_iter=1000, n_init=6, random_state=0)\n",
    "Last_seen_clusters = model.fit_predict(X)\n",
    "Last_seen_clusters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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>Weeks_since_last_seen</th>\n",
       "      <th>Last_seen_cluster</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Candidate_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>46</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>116</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Weeks_since_last_seen  Last_seen_cluster\n",
       "Candidate_id                                          \n",
       "1                                46                  2\n",
       "2                                 0                  0\n",
       "3                                 7                  0\n",
       "4                               116                  5\n",
       "5                                 1                  0"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Last_seen_clusters = pd.DataFrame(Last_seen_clusters,index = df_std.index,columns=['Last_seen_cluster'])\n",
    "\n",
    "df_pred = df.join(Last_seen_clusters)\n",
    "df_pred.fillna(0,inplace=True)\n",
    "df_pred.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Last_seen_cluster\n",
       "0      0\n",
       "1    150\n",
       "2     23\n",
       "3     58\n",
       "4    210\n",
       "5    100\n",
       "Name: Weeks_since_last_seen, dtype: int64"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pred.groupby(['Last_seen_cluster'])['Weeks_since_last_seen'].min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Last_seen_cluster\n",
       "0     22\n",
       "1    206\n",
       "2     57\n",
       "3     99\n",
       "4    373\n",
       "5    149\n",
       "Name: Weeks_since_last_seen, dtype: int64"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pred.groupby(['Last_seen_cluster'])['Weeks_since_last_seen'].max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_pred.loc[(df_pred['Weeks_since_last_seen'] >= 0 ) & (df_pred['Weeks_since_last_seen'] <= 22 ) , 'last_seen_weeks'] = '0 - 22'\n",
    "df_pred.loc[(df_pred['Weeks_since_last_seen'] >= 23 ) & (df_pred['Weeks_since_last_seen'] <= 56 ) , 'last_seen_weeks'] = '23 - 56'\n",
    "df_pred.loc[(df_pred['Weeks_since_last_seen'] >= 57 ) & (df_pred['Weeks_since_last_seen'] <= 99 ) , 'last_seen_weeks'] = '57 - 99'\n",
    "df_pred.loc[(df_pred['Weeks_since_last_seen'] >= 100 ) & (df_pred['Weeks_since_last_seen'] <= 150 ) , 'last_seen_weeks'] = '100 - 150'\n",
    "df_pred.loc[(df_pred['Weeks_since_last_seen'] >= 151 ) & (df_pred['Weeks_since_last_seen'] <= 206 ) , 'last_seen_weeks'] = '151 - 206'\n",
    "df_pred.loc[(df_pred['Weeks_since_last_seen'] >= 210 ) & (df_pred['Weeks_since_last_seen'] <= 373 ) , 'last_seen_weeks'] = '210 - 373'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Weeks_since_last_seen    1960\n",
       "Last_seen_cluster        1960\n",
       "last_seen_weeks          1960\n",
       "dtype: int64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pred.head()\n",
    "df_pred.count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# df_pred.reset_index(inplace=True,drop=True)\n",
    "# colors = ['#DF2020', '#81DF20', '#2095DF','#A985AF','#F095DF','A095DF']\n",
    "# df_pred['c'] = df_pred['Last_seen_cluster'].map({0:colors[0], 1:colors[1], 2:colors[2],3:colors[3], 4:colors[4], 5:colors[5]})\n",
    "# plt.scatter(df_pred['Last_seen_cluster'], df_pred['Weeks_since_last_seen'], c=df_pred['c'].values, alpha = 0.6, s=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<code style=\"background:yellow;color:black\"><b>Customers who were last seen between 0-22 weeks are most likely to be open for job openings </b></code> "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 3 : Case 2 (Predict Likelihood of Candidates being open to A SPECIFIC Job postings using KPrototypes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-info\">\n",
    "<b>Making predictions basis all relevant information including JOB ID since the prediction need to be Job specific </b> </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-32-84f4b915edd9>:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  df.dropna(inplace=True)\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></th>\n",
       "      <th>Exp_required</th>\n",
       "      <th>Candidate_exp</th>\n",
       "      <th>Weeks_since_last_seen</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Candidate_id</th>\n",
       "      <th>Job_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>979</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>18</td>\n",
       "      <td>115</td>\n",
       "      <td>0.606158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>8</td>\n",
       "      <td>75</td>\n",
       "      <td>0.492805</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>293</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "      <td>74</td>\n",
       "      <td>0.373916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1896</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>0.364500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>898</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>14</td>\n",
       "      <td>74</td>\n",
       "      <td>0.349587</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Exp_required  Candidate_exp  Weeks_since_last_seen  \\\n",
       "Candidate_id Job_id                                                       \n",
       "979          1                  6             18                    115   \n",
       "71           1                  6              8                     75   \n",
       "293          1                  6              7                     74   \n",
       "1896         1                  6              6                      0   \n",
       "898          1                  6             14                     74   \n",
       "\n",
       "                        score  \n",
       "Candidate_id Job_id            \n",
       "979          1       0.606158  \n",
       "71           1       0.492805  \n",
       "293          1       0.373916  \n",
       "1896         1       0.364500  \n",
       "898          1       0.349587  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = Final_df[['Candidate_id','Job_id','Exp_required','Candidate_exp','Weeks_since_last_seen','score']]\n",
    "df.dropna(inplace=True)\n",
    "df = df.set_index(['Candidate_id','Job_id'])\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Outlier treatment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "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></th>\n",
       "      <th>Exp_required</th>\n",
       "      <th>Candidate_exp</th>\n",
       "      <th>Weeks_since_last_seen</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Candidate_id</th>\n",
       "      <th>Job_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>979</th>\n",
       "      <th>1</th>\n",
       "      <td>1.167581</td>\n",
       "      <td>2.663652</td>\n",
       "      <td>0.852387</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <th>1</th>\n",
       "      <td>1.167581</td>\n",
       "      <td>0.755899</td>\n",
       "      <td>0.234533</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>293</th>\n",
       "      <th>1</th>\n",
       "      <td>1.167581</td>\n",
       "      <td>0.565123</td>\n",
       "      <td>0.219086</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1896</th>\n",
       "      <th>1</th>\n",
       "      <td>1.167581</td>\n",
       "      <td>0.374348</td>\n",
       "      <td>-0.923945</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>898</th>\n",
       "      <th>1</th>\n",
       "      <td>1.167581</td>\n",
       "      <td>1.900551</td>\n",
       "      <td>0.219086</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Exp_required  Candidate_exp  Weeks_since_last_seen  score\n",
       "Candidate_id Job_id                                                           \n",
       "979          1           1.167581       2.663652               0.852387    3.0\n",
       "71           1           1.167581       0.755899               0.234533    3.0\n",
       "293          1           1.167581       0.565123               0.219086    3.0\n",
       "1896         1           1.167581       0.374348              -0.923945    3.0\n",
       "898          1           1.167581       1.900551               0.219086    3.0"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Missing values\n",
    "df = df.fillna(0)\n",
    "df = df.replace([np.inf, -np.inf], np.nan)\n",
    "df.isnull().values.any()\n",
    "\n",
    "#standardize\n",
    "#df_std = pd.DataFrame(StandardScaler().fit_transform(df))\n",
    "\n",
    "df_std = pd.DataFrame(StandardScaler().fit_transform(df), index=df.index, columns=df.columns)\n",
    "\n",
    "#Treat outliers (Removing obs that fall beyond 3 stand dev)\n",
    "for col in df_std.columns:\n",
    "    upper = 3\n",
    "    lower = -3\n",
    "    df_std[col][df_std[col] < lower] = lower\n",
    "    df_std[col][df_std[col] > upper] = upper\n",
    "\n",
    "df_std.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>Job_id</th>\n",
       "      <th>Exp_required</th>\n",
       "      <th>Candidate_exp</th>\n",
       "      <th>Weeks_since_last_seen</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Candidate_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>979</th>\n",
       "      <td>1</td>\n",
       "      <td>1.167581</td>\n",
       "      <td>2.663652</td>\n",
       "      <td>0.852387</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <td>1</td>\n",
       "      <td>1.167581</td>\n",
       "      <td>0.755899</td>\n",
       "      <td>0.234533</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>293</th>\n",
       "      <td>1</td>\n",
       "      <td>1.167581</td>\n",
       "      <td>0.565123</td>\n",
       "      <td>0.219086</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1896</th>\n",
       "      <td>1</td>\n",
       "      <td>1.167581</td>\n",
       "      <td>0.374348</td>\n",
       "      <td>-0.923945</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>898</th>\n",
       "      <td>1</td>\n",
       "      <td>1.167581</td>\n",
       "      <td>1.900551</td>\n",
       "      <td>0.219086</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Job_id  Exp_required  Candidate_exp  Weeks_since_last_seen  score\n",
       "Candidate_id                                                                  \n",
       "979               1      1.167581       2.663652               0.852387    3.0\n",
       "71                1      1.167581       0.755899               0.234533    3.0\n",
       "293               1      1.167581       0.565123               0.219086    3.0\n",
       "1896              1      1.167581       0.374348              -0.923945    3.0\n",
       "898               1      1.167581       1.900551               0.219086    3.0"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2 = df_std.reset_index()\n",
    "df2.set_index('Candidate_id',inplace=True)\n",
    "# df2.head()\n",
    "df2['Job_id'] = df2['Job_id'].astype(str)\n",
    "df2.head()\n",
    "#df2.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([['1', 1.167580614787185, 2.663652124914918, 0.8523874165749968,\n",
       "        3.0],\n",
       "       ['1', 1.167580614787185, 0.755898648033664, 0.23453263237122696,\n",
       "        3.0],\n",
       "       ['1', 1.167580614787185, 0.5651233003455387, 0.21908626276613272,\n",
       "        3.0],\n",
       "       ...,\n",
       "       ['161', -1.2993548049075236, -0.7703041334713389,\n",
       "        0.049176197110095986, -0.2674865409262955],\n",
       "       ['161', -1.2993548049075236, 0.3743479526574133,\n",
       "        0.6052455028934889, -0.2674865409262955],\n",
       "       ['161', -1.2993548049075236, -0.5795287857832134,\n",
       "        0.35810358921198093, -0.2674865409262955]], dtype=object)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X=df2.values\n",
    "X"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Kprototype for obtaining Clusters"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Decide optimal clusters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "categorical_index = [0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # Function for plotting elbow curve\n",
    "# def plot_elbow_curve(start, end, data):\n",
    "#     no_of_clusters = list(range(start, end+1))\n",
    "#     cost_values = []\n",
    "    \n",
    "#     for k in no_of_clusters:\n",
    "#         test_model = KPrototypes(n_clusters=k, init='Huang', random_state=42)\n",
    "#         test_model.fit_predict(data, categorical=categorical_index)\n",
    "#         cost_values.append(test_model.cost_)\n",
    "        \n",
    "#     sns.set_theme(style=\"whitegrid\", palette=\"bright\", font_scale=1.2)\n",
    "    \n",
    "#     plt.figure(figsize=(15, 7))\n",
    "#     ax = sns.lineplot(x=no_of_clusters, y=cost_values, marker=\"o\", dashes=False)\n",
    "#     ax.set_title('Elbow curve', fontsize=18)\n",
    "#     ax.set_xlabel('No of clusters', fontsize=14)\n",
    "#     ax.set_ylabel('Cost', fontsize=14)\n",
    "#     ax.set(xlim=(start-0.1, end+0.1))\n",
    "#     plt.plot();\n",
    "    \n",
    "# # Plotting elbow curve for k=2 to k=10\n",
    "# plot_elbow_curve(4,10,X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best run was number 4\n",
      "510991.55439440766\n",
      "--- 973.3187763690948 seconds ---\n"
     ]
    }
   ],
   "source": [
    "start_time = time.time()\n",
    "start_time\n",
    "\n",
    "model = KPrototypes(n_clusters=5, init='Huang', random_state=42, n_jobs=-1,verbose=True)\n",
    "model.fit_predict(X, categorical=categorical_index)\n",
    "print(model.cost_)\n",
    "\n",
    "print(\"--- %s seconds ---\" % (time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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></th>\n",
       "      <th>Exp_required</th>\n",
       "      <th>Candidate_exp</th>\n",
       "      <th>Weeks_since_last_seen</th>\n",
       "      <th>score</th>\n",
       "      <th>clusters</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Candidate_id</th>\n",
       "      <th>Job_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>979</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>18</td>\n",
       "      <td>115</td>\n",
       "      <td>0.606158</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>8</td>\n",
       "      <td>75</td>\n",
       "      <td>0.492805</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>293</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "      <td>74</td>\n",
       "      <td>0.373916</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1896</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>0.364500</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>898</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>14</td>\n",
       "      <td>74</td>\n",
       "      <td>0.349587</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Exp_required  Candidate_exp  Weeks_since_last_seen  \\\n",
       "Candidate_id Job_id                                                       \n",
       "979          1                  6             18                    115   \n",
       "71           1                  6              8                     75   \n",
       "293          1                  6              7                     74   \n",
       "1896         1                  6              6                      0   \n",
       "898          1                  6             14                     74   \n",
       "\n",
       "                        score  clusters  \n",
       "Candidate_id Job_id                      \n",
       "979          1       0.606158         4  \n",
       "71           1       0.492805         4  \n",
       "293          1       0.373916         4  \n",
       "1896         1       0.364500         4  \n",
       "898          1       0.349587         4  "
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clusters = pd.DataFrame(model.labels_,index = df_std.index,columns=['clusters'])\n",
    "\n",
    "df_pred = df.join(clusters)\n",
    "df_pred.fillna(0,inplace=True)\n",
    "df_pred.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-40-5fa42b9b3b58>:1: FutureWarning: Indexing with multiple keys (implicitly converted to a tuple of keys) will be deprecated, use a list instead.\n",
      "  df_pred.groupby(['clusters'])['Exp_required','Candidate_exp','Weeks_since_last_seen','score'].min()\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>Exp_required</th>\n",
       "      <th>Candidate_exp</th>\n",
       "      <th>Weeks_since_last_seen</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>clusters</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>86</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.057346</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Exp_required  Candidate_exp  Weeks_since_last_seen     score\n",
       "clusters                                                              \n",
       "0                    1              0                      0  0.000000\n",
       "1                    1              7                      0  0.000000\n",
       "2                    1              0                     86  0.000000\n",
       "3                    5              0                      0  0.000000\n",
       "4                    1              0                      0  0.057346"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pred.groupby(['clusters'])['Exp_required','Candidate_exp','Weeks_since_last_seen','score'].min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_pred.reset_index(inplace=True)\n",
    "colors = ['#DF2020', '#81DF20', '#2095DF','#A985AF','#F095DF']\n",
    "df_pred2 = df_pred\n",
    "df_pred2['c'] = df_pred.clusters.map({0:colors[0], 1:colors[1], 2:colors[2],3:colors[3], 4:colors[4]})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fb8e7d1bb20>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "plt.scatter(df_pred2['clusters'], df_pred2['Exp_required'], c=df_pred2.c, alpha = 0.6, s=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fb8e7cc11f0>"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "plt.scatter(df_pred2['clusters'], df_pred2['Candidate_exp'], c=df_pred2.c, alpha = 0.6, s=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fb8e7bf2520>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "plt.scatter(df_pred2['clusters'], df_pred2['Weeks_since_last_seen'], c=df_pred2.c, alpha = 0.6, s=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fb8e7a01340>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "plt.scatter(df_pred2['clusters'], df_pred2['score'], c=df_pred2.c, alpha = 0.6, s=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<code style=\"background:yellow;color:black\"><b>Inference </b></code> "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Cluster 4 seems to be the only cluster that has  higher Job vs Candidate match (>.2)\n",
    "- We can candidates from cluster 4 where score > .5 and last_seen < 4 (they were last seen within a month) and whose experience is somewhere close to required Job experience"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "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>Candidate_id</th>\n",
       "      <th>Job_id</th>\n",
       "      <th>Exp_required</th>\n",
       "      <th>Candidate_exp</th>\n",
       "      <th>Weeks_since_last_seen</th>\n",
       "      <th>score</th>\n",
       "      <th>clusters</th>\n",
       "      <th>c</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3920</th>\n",
       "      <td>1687</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0.595328</td>\n",
       "      <td>4</td>\n",
       "      <td>#F095DF</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5880</th>\n",
       "      <td>174</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0.752767</td>\n",
       "      <td>4</td>\n",
       "      <td>#F095DF</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7840</th>\n",
       "      <td>1146</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.565995</td>\n",
       "      <td>4</td>\n",
       "      <td>#F095DF</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17640</th>\n",
       "      <td>1207</td>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.890725</td>\n",
       "      <td>4</td>\n",
       "      <td>#F095DF</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19600</th>\n",
       "      <td>251</td>\n",
       "      <td>11</td>\n",
       "      <td>2</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "      <td>0.778986</td>\n",
       "      <td>4</td>\n",
       "      <td>#F095DF</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Candidate_id  Job_id  Exp_required  Candidate_exp  \\\n",
       "3920           1687       3             2              7   \n",
       "5880            174       4             2              2   \n",
       "7840           1146       5             1              0   \n",
       "17640          1207      10             2              1   \n",
       "19600           251      11             2              9   \n",
       "\n",
       "       Weeks_since_last_seen     score  clusters        c  \n",
       "3920                       0  0.595328         4  #F095DF  \n",
       "5880                       2  0.752767         4  #F095DF  \n",
       "7840                       0  0.565995         4  #F095DF  \n",
       "17640                      0  0.890725         4  #F095DF  \n",
       "19600                      0  0.778986         4  #F095DF  "
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pred2[(df_pred2['clusters'] == 4) & (df_pred2['Weeks_since_last_seen'] <=4) & (df_pred2['score'] >= .5)].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": 2
}