2_rank-postings.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "0613b014",
   "metadata": {},
   "source": [
    "<span style=\"font-size: 120%\">Limits the job postings dataset to those most similar to the user's resume allowing:</span>\n",
    " - <span style=\"font-size: 120%\">faster k-means processing in the following notebook</span>\n",
    " - <span style=\"font-size: 120%\">less individual jobs postings to sift through from the\n",
    "   perspective of a job seeker</span>\n",
    "\n",
    "<span style=\"font-size: 120%\">However doing so will almost certainly change the sample\n",
    "from a statistical point of view.<br>That is, make it less representative of the \n",
    "original population, and more representative of <br>a different population given the sample\n",
    "was truncated non-arbitrarily.</span>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "eff2e12a",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics.pairwise        import cosine_similarity\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "\n",
    "%matplotlib inline\n",
    "from matplotlib import pyplot as plt\n",
    "import numpy  as np\n",
    "import pandas as pd\n",
    "\n",
    "import PyPDF2 as pp2\n",
    "\n",
    "import json\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "32bf7e15",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 1)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ngram_param = None\n",
    "with open('config.json') as f:\n",
    "    ngram_param = tuple(json.load(f)['ngram_range'])\n",
    "ngram_param"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d36c50e0",
   "metadata": {},
   "outputs": [],
   "source": [
    "resume_dict = None\n",
    "with open('data/Resume_Tek-Cub.pdf', 'rb') as f:\n",
    "    page = pp2.PdfFileReader(f).getPage(0)\n",
    "    resume_dict = {'title':['resume'], 'body':[page.extractText()], 'li_elems':['']}\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "b96febe8",
   "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>title</th>\n",
       "      <th>body</th>\n",
       "      <th>li_elems</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Senior Data Scientist - Risk Mitigation - Jers...</td>\n",
       "      <td>Senior Data Scientist - Risk Mitigation - Jers...</td>\n",
       "      <td>(3+ years of experience crafting, developing, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Director, Data Scientist, Natural Language Pro...</td>\n",
       "      <td>Director, Data Scientist, Natural Language Pro...</td>\n",
       "      <td>(Lead workshops, innovation sessions with clie...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>eCommerce Senior Data Scientist - United States</td>\n",
       "      <td>eCommerce Senior Data Scientist - United State...</td>\n",
       "      <td>(Build machine learning systems to understand ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Associate Scientist, Data Science - Irving, TX...</td>\n",
       "      <td>Associate Scientist, Data Science - Irving, TX...</td>\n",
       "      <td>(Research and development of new products., Pr...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Data Scientist - Seattle, WA</td>\n",
       "      <td>Data Scientist - Seattle, WA\\nA Bachelor or Ma...</td>\n",
       "      <td>(A Bachelor or Masters Degree in a highly quan...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                title  \\\n",
       "3   Senior Data Scientist - Risk Mitigation - Jers...   \n",
       "9   Director, Data Scientist, Natural Language Pro...   \n",
       "10    eCommerce Senior Data Scientist - United States   \n",
       "11  Associate Scientist, Data Science - Irving, TX...   \n",
       "12                       Data Scientist - Seattle, WA   \n",
       "\n",
       "                                                 body  \\\n",
       "3   Senior Data Scientist - Risk Mitigation - Jers...   \n",
       "9   Director, Data Scientist, Natural Language Pro...   \n",
       "10  eCommerce Senior Data Scientist - United State...   \n",
       "11  Associate Scientist, Data Science - Irving, TX...   \n",
       "12  Data Scientist - Seattle, WA\\nA Bachelor or Ma...   \n",
       "\n",
       "                                             li_elems  \n",
       "3   (3+ years of experience crafting, developing, ...  \n",
       "9   (Lead workshops, innovation sessions with clie...  \n",
       "10  (Build machine learning systems to understand ...  \n",
       "11  (Research and development of new products., Pr...  \n",
       "12  (A Bachelor or Masters Degree in a highly quan...  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "jobs_df = pd.read_pickle('data/jobs_df.pkl', 'zip')\n",
    "jobs_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "02ebd5ed",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(493, 3)\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>title</th>\n",
       "      <th>body</th>\n",
       "      <th>li_elems</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>resume</td>\n",
       "      <td>Tek Cub\\nFAMILIAR TECH / CONCEPTS\\nLanguages: ...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Senior Data Scientist - Risk Mitigation - Jers...</td>\n",
       "      <td>Senior Data Scientist - Risk Mitigation - Jers...</td>\n",
       "      <td>(3+ years of experience crafting, developing, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Director, Data Scientist, Natural Language Pro...</td>\n",
       "      <td>Director, Data Scientist, Natural Language Pro...</td>\n",
       "      <td>(Lead workshops, innovation sessions with clie...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>eCommerce Senior Data Scientist - United States</td>\n",
       "      <td>eCommerce Senior Data Scientist - United State...</td>\n",
       "      <td>(Build machine learning systems to understand ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Associate Scientist, Data Science - Irving, TX...</td>\n",
       "      <td>Associate Scientist, Data Science - Irving, TX...</td>\n",
       "      <td>(Research and development of new products., Pr...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                title  \\\n",
       "0                                              resume   \n",
       "3   Senior Data Scientist - Risk Mitigation - Jers...   \n",
       "9   Director, Data Scientist, Natural Language Pro...   \n",
       "10    eCommerce Senior Data Scientist - United States   \n",
       "11  Associate Scientist, Data Science - Irving, TX...   \n",
       "\n",
       "                                                 body  \\\n",
       "0   Tek Cub\\nFAMILIAR TECH / CONCEPTS\\nLanguages: ...   \n",
       "3   Senior Data Scientist - Risk Mitigation - Jers...   \n",
       "9   Director, Data Scientist, Natural Language Pro...   \n",
       "10  eCommerce Senior Data Scientist - United State...   \n",
       "11  Associate Scientist, Data Science - Irving, TX...   \n",
       "\n",
       "                                             li_elems  \n",
       "0                                                      \n",
       "3   (3+ years of experience crafting, developing, ...  \n",
       "9   (Lead workshops, innovation sessions with clie...  \n",
       "10  (Build machine learning systems to understand ...  \n",
       "11  (Research and development of new products., Pr...  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# complete the dataset by including the user's resume document \n",
    "\n",
    "dataset_df = pd.DataFrame(resume_dict).append(jobs_df)\n",
    "print(dataset_df.shape)\n",
    "dataset_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "88fcb62b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# begin using NLP techniques:\n",
    "# - use tf-idf as embedding: transform resume and dataset into tf-idf vectors\n",
    "# - use unigrams, bi-grams, ansd tri-grams\n",
    " \n",
    "# use content of job posting webpages and resume as corpus to derive vocabulary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "7f24a5e0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "resume vector dimensions: (1, 10035)\n",
      "jobs vector dimensions:   (492, 10035)\n"
     ]
    }
   ],
   "source": [
    "corpus = dataset_df['body']\n",
    "vocab  = TfidfVectorizer(ngram_range=ngram_param, stop_words='english')\n",
    "freq_matrix = vocab.fit_transform(corpus)\n",
    "\n",
    "resume_vec = freq_matrix[0, :]\n",
    "jobs_vec = freq_matrix[1:, :]\n",
    "\n",
    "print(f\"resume vector dimensions: {resume_vec.shape}\")\n",
    "print(f\"jobs vector dimensions:   {jobs_vec.shape}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "6a32c090",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(492, 1)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# see how similar resume is to the rest of dataset\n",
    "cos_sims = cosine_similarity(jobs_vec, resume_vec)\n",
    "cos_sims.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "9cf0f73e",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.05314223398462385    0.08922963828567279    0.046371434627420036    0.11043520680240497    "
     ]
    }
   ],
   "source": [
    "# check some elements for expected results\n",
    "for i in range(4):\n",
    "    print(cos_sims[i][0], end='    ');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "05e8b21e",
   "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>title</th>\n",
       "      <th>body</th>\n",
       "      <th>li_elems</th>\n",
       "      <th>cosine_similarity</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Data Scientist - Chesterfield, MO</td>\n",
       "      <td>Data Scientist - Chesterfield, MO\\nCompany Det...</td>\n",
       "      <td>(Take high-level project requirements and form...</td>\n",
       "      <td>0.127448</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Data Scientist - Chesterfield, MO 63017</td>\n",
       "      <td>Data Scientist - Chesterfield, MO 63017\\nCompa...</td>\n",
       "      <td>(Take high-level project requirements and form...</td>\n",
       "      <td>0.127190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Data Science Software Engineer - San Diego, CA...</td>\n",
       "      <td>Data Science Software Engineer - San Diego, CA...</td>\n",
       "      <td>(Organize, transform, and verify the integrity...</td>\n",
       "      <td>0.126004</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Data Scientist - San Francisco, CA</td>\n",
       "      <td>Data Scientist - San Francisco, CA\\nJob Descri...</td>\n",
       "      <td>(Perform hands-on analysis of large volumes of...</td>\n",
       "      <td>0.119804</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Senior Data Scientist - San Francisco, CA</td>\n",
       "      <td>Senior Data Scientist - San Francisco, CA\\nJob...</td>\n",
       "      <td>(Perform hands-on analysis of large volumes of...</td>\n",
       "      <td>0.119635</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>487</th>\n",
       "      <td>Senior Data Scientist - San Francisco, CA 94111</td>\n",
       "      <td>Senior Data Scientist - San Francisco, CA 9411...</td>\n",
       "      <td>(Be a subject matter expert on all things rela...</td>\n",
       "      <td>0.030222</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>488</th>\n",
       "      <td>Post-doc: Dissemination/implementation and/or ...</td>\n",
       "      <td>Post-doc: Dissemination/implementation and/or ...</td>\n",
       "      <td>()</td>\n",
       "      <td>0.028809</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>489</th>\n",
       "      <td>Data Scientist - West Lafayette, IN 47906</td>\n",
       "      <td>Data Scientist - West Lafayette, IN 47906\\nDTN...</td>\n",
       "      <td>(Gain experience in all areas of data science,...</td>\n",
       "      <td>0.028500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>490</th>\n",
       "      <td>Professor of Data Sciences and Operations (Ope...</td>\n",
       "      <td>Professor of Data Sciences and Operations (Ope...</td>\n",
       "      <td>()</td>\n",
       "      <td>0.019954</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>491</th>\n",
       "      <td>Part-time instructor of Statistics for Data Sc...</td>\n",
       "      <td>Part-time instructor of Statistics for Data Sc...</td>\n",
       "      <td>()</td>\n",
       "      <td>0.013871</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>492 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 title  \\\n",
       "0                    Data Scientist - Chesterfield, MO   \n",
       "1              Data Scientist - Chesterfield, MO 63017   \n",
       "2    Data Science Software Engineer - San Diego, CA...   \n",
       "3                   Data Scientist - San Francisco, CA   \n",
       "4            Senior Data Scientist - San Francisco, CA   \n",
       "..                                                 ...   \n",
       "487    Senior Data Scientist - San Francisco, CA 94111   \n",
       "488  Post-doc: Dissemination/implementation and/or ...   \n",
       "489          Data Scientist - West Lafayette, IN 47906   \n",
       "490  Professor of Data Sciences and Operations (Ope...   \n",
       "491  Part-time instructor of Statistics for Data Sc...   \n",
       "\n",
       "                                                  body  \\\n",
       "0    Data Scientist - Chesterfield, MO\\nCompany Det...   \n",
       "1    Data Scientist - Chesterfield, MO 63017\\nCompa...   \n",
       "2    Data Science Software Engineer - San Diego, CA...   \n",
       "3    Data Scientist - San Francisco, CA\\nJob Descri...   \n",
       "4    Senior Data Scientist - San Francisco, CA\\nJob...   \n",
       "..                                                 ...   \n",
       "487  Senior Data Scientist - San Francisco, CA 9411...   \n",
       "488  Post-doc: Dissemination/implementation and/or ...   \n",
       "489  Data Scientist - West Lafayette, IN 47906\\nDTN...   \n",
       "490  Professor of Data Sciences and Operations (Ope...   \n",
       "491  Part-time instructor of Statistics for Data Sc...   \n",
       "\n",
       "                                              li_elems  cosine_similarity  \n",
       "0    (Take high-level project requirements and form...           0.127448  \n",
       "1    (Take high-level project requirements and form...           0.127190  \n",
       "2    (Organize, transform, and verify the integrity...           0.126004  \n",
       "3    (Perform hands-on analysis of large volumes of...           0.119804  \n",
       "4    (Perform hands-on analysis of large volumes of...           0.119635  \n",
       "..                                                 ...                ...  \n",
       "487  (Be a subject matter expert on all things rela...           0.030222  \n",
       "488                                                 ()           0.028809  \n",
       "489  (Gain experience in all areas of data science,...           0.028500  \n",
       "490                                                 ()           0.019954  \n",
       "491                                                 ()           0.013871  \n",
       "\n",
       "[492 rows x 4 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# match job postings with their respective similarities\n",
    "jobs_df['cosine_similarity'] = cos_sims\n",
    "sorted_jobs_df = jobs_df.sort_values(by='cosine_similarity', ascending=False)\n",
    "sorted_jobs_df = sorted_jobs_df.reset_index(drop=True)\n",
    "sorted_jobs_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "12a2a24b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10, 6))\n",
    "plt.title('Cosine Similarity of Job Postings to Resume')\n",
    "plt.xlabel('job postings quantity')\n",
    "plt.ylabel('cosine similarity')\n",
    "plt.axvline(x=80, color='purple')\n",
    "\n",
    "sorted_jobs_df['cosine_similarity'].plot();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "4d36e0bc",
   "metadata": {},
   "outputs": [],
   "source": [
    "# intuiting from chart's slope: save the first 80 most similar job postings\n",
    "pd.DataFrame(sorted_jobs_df[:80]).to_pickle('data/sorted_trunc_jobs_df.pkl', 'zip')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.6.9"
  },
  "vscode": {
   "interpreter": {
    "hash": "31f2aee4e71d21fbe5cf8b01ff0e069b9275f58929596ceb00d14d90e3e16cd6"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}