collect_literature_updated.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import io\n",
    "import os\n",
    "import xml.etree.ElementTree as et \n",
    "import xmltodict\n",
    "import urllib\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "\n",
    "os.environ[\"PATH\"] += os.pathsep + 'C:\\\\Program Files\\\\MiKTeX 2.9\\\\miktex\\\\bin\\\\x64\\\\'\n",
    "os.environ[\"PATH\"] += os.pathsep + 'C:\\\\Program Files\\\\gs\\\\gs9.52\\\\bin\\\\'\n",
    "matplotlib.rcParams['text.usetex'] = True"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Get data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "COVID-19 Open Research Dataset (CORD-19). 2020. Version 2020-08-04. Retrieved from https://ai2-semanticscholar-cord-19.s3-us-west-2.amazonaws.com/historical_releases.html. Accessed 2020-08-04. doi:10.5281/zenodo.3715505\n",
    "\n",
    "The query used by the authors is as follows:\n",
    "\n",
    "    \"COVID-19\"[All Fields] OR (\"coronavirus\"[MeSH Terms] OR \"coronavirus\"[All Fields]) OR \"Corona virus\"[All Fields] OR \"2019-nCoV\"[All Fields] OR \"SARS-CoV\"[All Fields] OR \"MERS-CoV\"[All Fields] OR \"Severe Acute Respiratory Syndrome\"[All Fields] OR \"Middle East Respiratory Syndrome\"[All Fields]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(207595, 19)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cord19 = pd.read_csv(\"https://ai2-semanticscholar-cord-19.s3-us-west-2.amazonaws.com/latest/metadata.csv\", low_memory=False) # last retrieved: 2020-05-01\n",
    "cord19.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(125428, 22)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cord19.rename(columns = {'cord_uid'      :'id', \n",
    "                          'publish_time' :'published', \n",
    "                          'abstract'     :'summary', \n",
    "                          'authors'      :'author',  \n",
    "                          'source_x'     :'source'}, inplace=True)\n",
    "\n",
    "# Extract mentions of individual ArXiv sources\n",
    "cord19['source'] = cord19['source'].str.lower()\n",
    "cord19['medrxiv'] = cord19['source'].str.contains('medrxiv')\n",
    "cord19['arxiv'  ] = cord19['source'].str.contains('arxiv')\n",
    "cord19['biorxiv'] = cord19['source'].str.contains('biorxiv')\n",
    "\n",
    "# Keep only articles with title and summary, and only articles published in 2019 or 2020\n",
    "cord19.dropna(subset=['title', 'summary'], how='all', inplace=True) \n",
    "cord19['published'] = pd.to_datetime(cord19['published'])\n",
    "cord19 = cord19[cord19.published.dt.year.isin([2019,2020])].copy()\n",
    "\n",
    "cord19.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Extract AI papers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "cord19['paper_type'] = 'Not AI'\n",
    "cord19['paper_type'] = np.where(cord19['title'  ].astype(str).str.lower().str.contains(\"(machine learning|artificial intelligence|deep learning|neural network)\"), \n",
    "                                'AI', \n",
    "                                cord19['paper_type'])\n",
    "cord19['paper_type'] = np.where(cord19['summary'].astype(str).str.lower().str.contains(\"(machine learning|artificial intelligence|deep learning|neural network)\"), \n",
    "                                'AI', \n",
    "                                cord19['paper_type'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Remove duplicates"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Variable:       sha \n",
      "Max count: 3\n",
      "True duplicates 7\n",
      "N dropped       7 \n",
      "\n",
      "Variable:       doi \n",
      "Max count: 2\n",
      "True duplicates 2\n",
      "N dropped       2 \n",
      "\n",
      "Variable:       pubmed_id \n",
      "Max count: 3\n",
      "True duplicates 48\n",
      "N dropped       48 \n",
      "\n",
      "Variable:       url \n",
      "Max count: 1\n",
      "True duplicates 0\n",
      "N dropped       0 \n",
      "\n",
      "Variable:       s2_id \n",
      "Max count: 20\n",
      "True duplicates 80057\n",
      "N dropped       80057 \n",
      "\n",
      "Variable:       arxiv_id \n",
      "Max count: 1\n",
      "True duplicates 0\n",
      "N dropped       0 \n",
      "\n",
      "Variable:       pmcid \n",
      "Max count: 1\n",
      "True duplicates 0\n",
      "N dropped       0 \n",
      "\n",
      "Variable:       who_covidence_id \n",
      "Max count: 1\n",
      "True duplicates 0\n",
      "N dropped       0 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Naively, on ID\n",
    "cord19.drop_duplicates(subset=['id'], inplace=True)\n",
    "\n",
    "# On IDs pulled from other sources\n",
    "unique_identifiers = ['sha', 'doi',  'pubmed_id',  'url', 's2_id', 'arxiv_id', 'pmcid', 'who_covidence_id']\n",
    "\n",
    "for u in unique_identifiers:\n",
    "    \n",
    "    # Get the maximum value count of a single ID\n",
    "    print(\"Variable:      \", u, \"\\nMax count:\", cord19[u].value_counts().max())\n",
    "    \n",
    "    # Figure out how many true duplicates there are (ignoring NAN counted as duplicate)\n",
    "    missing    = cord19[u].isna().value_counts().loc[True]\n",
    "    duplicated = cord19[u].duplicated(keep=False).value_counts().loc[True]\n",
    "\n",
    "    print(\"True duplicates\", duplicated - missing)\n",
    "    \n",
    "    # Count how many records drop\n",
    "    shape_before = cord19.shape[0]    \n",
    "    cord19 = cord19[~(~cord19[u].isna() & cord19[u].duplicated(keep=False))].copy()\n",
    "    shape_after = cord19.shape[0]\n",
    "    \n",
    "    print(\"N dropped      \", shape_before - shape_after, \"\\n\")\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Save all papers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1375, 23)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cord19[cord19.paper_type==\"AI\"].to_csv(\"list_of_papers.csv\", encoding='utf-8-sig')\n",
    "cord19[cord19.paper_type==\"AI\"].shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Analyze"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Total counts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1305"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# How many AI papers in our time window?\n",
    "cord19[cord19.paper_type==\"AI\"].set_index(\"published\").loc['2020-01-01':'2020-08-01'].count()['title']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "36292"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# How many total COVID papers in our time window?\n",
    "cord19.set_index(\"published\").loc['2020-01-01':'2020-08-01'].count()['title']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "pmc                                   10981\n",
       "medline                                5493\n",
       "medrxiv                                5128\n",
       "medline; pmc                           4375\n",
       "who                                    3933\n",
       "arxiv                                  2252\n",
       "elsevier; pmc; who                     2181\n",
       "elsevier; medline; pmc                 1916\n",
       "pmc; who                               1560\n",
       "biorxiv                                1425\n",
       "medline; pmc; who                      1388\n",
       "elsevier; pmc                          1187\n",
       "elsevier; medline; pmc; who             558\n",
       "medline; who                            507\n",
       "biorxiv; medrxiv                        197\n",
       "medrxiv; medline; pmc                   192\n",
       "elsevier                                144\n",
       "medrxiv; medline                        115\n",
       "biorxiv; medline; pmc                    77\n",
       "biorxiv; medline                         58\n",
       "biorxiv; medline; pmc; who               48\n",
       "biorxiv; medline; who                    23\n",
       "medrxiv; medline; pmc; who               20\n",
       "arxiv; elsevier; pmc; who                14\n",
       "medrxiv; pmc                             13\n",
       "medrxiv; medline; who                    10\n",
       "biorxiv; pmc                             10\n",
       "arxiv; medline; pmc                       8\n",
       "biorxiv; medrxiv; medline                 8\n",
       "biorxiv; medrxiv; medline; pmc            7\n",
       "arxiv; who                                6\n",
       "arxiv; medline                            6\n",
       "arxiv; elsevier; medline; pmc             5\n",
       "arxiv; elsevier; medline; pmc; who        5\n",
       "arxiv; elsevier; pmc                      5\n",
       "biorxiv; who                              5\n",
       "elsevier; who                             4\n",
       "medrxiv; who                              3\n",
       "arxiv; medline; pmc; who                  3\n",
       "arxiv; elsevier                           2\n",
       "arxiv; pmc; who                           1\n",
       "arxiv; medrxiv                            1\n",
       "arxiv; pmc                                1\n",
       "Name: source, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cord19.source.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1,2, figsize=[15,5])\n",
    "\n",
    "# Note: we start from 2 January because it looks like articles with missing date get assigned to Jan 1\n",
    "plot_df = cord19.set_index('published').loc['2020-01-02':'2020-08-04'].reset_index()\n",
    "\n",
    "###############################\n",
    "# Plot all papers (left panel)\n",
    "###############################\n",
    "\n",
    "plot_df1 = plot_df.groupby([plot_df.published.dt.to_period('W'), \n",
    "                            plot_df.paper_type]).count()['id'].unstack(level=['paper_type']).fillna(0)\n",
    "\n",
    "plot_df1['All papers'] = plot_df1['AI'] + plot_df1['Not AI']\n",
    "\n",
    "\n",
    "plot_df1[['AI', 'All papers']].loc['2020-01-02':'2020-04-10'].plot(ax=ax[0],  legend=True,  alpha=0.75, color=['blue', 'darkorange'], marker=\".\")\n",
    "plot_df1[['AI', 'All papers']].loc['2020-04-10':].plot(ax=ax[0],  legend=True,  alpha=1, color=['blue', 'darkorange'], linestyle=\":\")\n",
    "\n",
    "\n",
    "ax[0].set_xlabel(\"\")\n",
    "ax[0].set_ylabel(\"$\\#$ papers per week\")\n",
    "ax[0].set_title(\"All coronavirus papers\") #CORD-19 papers + ArXiv papers mentioning covid/coronavirus/sars-cov-2\")\n",
    "ax[0].legend([\"AI/ML-related papers\", \"All papers\"])\n",
    "\n",
    "##################################\n",
    "# Plot ML/AI papers (right panel)\n",
    "##################################\n",
    "\n",
    "plot_df2 = plot_df.groupby([plot_df.published.dt.to_period('W'),\n",
    "                            plot_df.paper_type]).sum()[['arxiv', 'medrxiv', 'biorxiv']].unstack(level=['paper_type']).fillna(0)\n",
    "\n",
    "plot_df2 = plot_df2.xs(\"AI\", level=\"paper_type\", axis=1).fillna(0)\n",
    "plot_df2[['arxiv', 'medrxiv', 'biorxiv']].loc['2020-01-02':'2020-04-10'].plot(alpha=0.75, ax=ax[1], cmap='gist_rainbow', marker=\".\", legend=True)\n",
    "plot_df2[['arxiv', 'medrxiv', 'biorxiv']].loc['2020-04-10':].plot(alpha=1, ax=ax[1], cmap='gist_rainbow', linestyle=\":\",  legend=False)\n",
    "\n",
    "ax[1].set_xlabel(\"\")\n",
    "ax[1].set_ylabel(\"$\\#$ papers per week\")\n",
    "ax[1].set_title(\"AI/ML-related papers posted on ArXiv, MedRxiv, and BioRxiv\") \n",
    "\n",
    "\n",
    "##################################\n",
    "# Polish and save plot\n",
    "##################################\n",
    "\n",
    "ax[0].set_xlim('2020-01-02', '2020-08-01')\n",
    "ax[1].set_xlim('2020-01-02', '2020-08-01')\n",
    "ax[0].set_ylim(-50, 2000)\n",
    "ax[1].set_ylim(-1, 35)\n",
    "\n",
    "plt.savefig(\"automated_lit_review.pdf\", dpi=100, bbox_inches=\"tight\")\n",
    "plt.savefig(\"automated_lit_review.png\", dpi=100, bbox_inches=\"tight\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "published\n",
       "2019-12-30/2020-01-05       79.0\n",
       "2020-01-06/2020-01-12      265.0\n",
       "2020-01-13/2020-01-19      404.0\n",
       "2020-01-20/2020-01-26      559.0\n",
       "2020-01-27/2020-02-02      801.0\n",
       "2020-02-03/2020-02-09     1005.0\n",
       "2020-02-10/2020-02-16     1247.0\n",
       "2020-02-17/2020-02-23     1569.0\n",
       "2020-02-24/2020-03-01     1907.0\n",
       "2020-03-02/2020-03-08     2273.0\n",
       "2020-03-09/2020-03-15     2677.0\n",
       "2020-03-16/2020-03-22     3202.0\n",
       "2020-03-23/2020-03-29     3877.0\n",
       "2020-03-30/2020-04-05     4824.0\n",
       "2020-04-06/2020-04-12     5808.0\n",
       "2020-04-13/2020-04-19     6929.0\n",
       "2020-04-20/2020-04-26     8105.0\n",
       "2020-04-27/2020-05-03     9488.0\n",
       "2020-05-04/2020-05-10    10993.0\n",
       "2020-05-11/2020-05-17    12588.0\n",
       "2020-05-18/2020-05-24    14371.0\n",
       "2020-05-25/2020-05-31    16261.0\n",
       "2020-06-01/2020-06-07    18087.0\n",
       "2020-06-08/2020-06-14    19949.0\n",
       "2020-06-15/2020-06-21    21650.0\n",
       "2020-06-22/2020-06-28    23430.0\n",
       "2020-06-29/2020-07-05    25226.0\n",
       "2020-07-06/2020-07-12    26791.0\n",
       "2020-07-13/2020-07-19    28415.0\n",
       "2020-07-20/2020-07-26    29982.0\n",
       "2020-07-27/2020-08-02    31764.0\n",
       "2020-08-03/2020-08-09    31857.0\n",
       "Freq: W-SUN, Name: All papers, dtype: float64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plot_df1['All papers'].cumsum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "published\n",
       "2019-12-30/2020-01-05       0.0\n",
       "2020-01-06/2020-01-12       2.0\n",
       "2020-01-13/2020-01-19       2.0\n",
       "2020-01-20/2020-01-26       3.0\n",
       "2020-01-27/2020-02-02      10.0\n",
       "2020-02-03/2020-02-09      14.0\n",
       "2020-02-10/2020-02-16      20.0\n",
       "2020-02-17/2020-02-23      25.0\n",
       "2020-02-24/2020-03-01      30.0\n",
       "2020-03-02/2020-03-08      36.0\n",
       "2020-03-09/2020-03-15      44.0\n",
       "2020-03-16/2020-03-22      74.0\n",
       "2020-03-23/2020-03-29     109.0\n",
       "2020-03-30/2020-04-05     141.0\n",
       "2020-04-06/2020-04-12     172.0\n",
       "2020-04-13/2020-04-19     244.0\n",
       "2020-04-20/2020-04-26     283.0\n",
       "2020-04-27/2020-05-03     349.0\n",
       "2020-05-04/2020-05-10     448.0\n",
       "2020-05-11/2020-05-17     498.0\n",
       "2020-05-18/2020-05-24     591.0\n",
       "2020-05-25/2020-05-31     694.0\n",
       "2020-06-01/2020-06-07     763.0\n",
       "2020-06-08/2020-06-14     879.0\n",
       "2020-06-15/2020-06-21     953.0\n",
       "2020-06-22/2020-06-28    1022.0\n",
       "2020-06-29/2020-07-05    1073.0\n",
       "2020-07-06/2020-07-12    1118.0\n",
       "2020-07-13/2020-07-19    1166.0\n",
       "2020-07-20/2020-07-26    1208.0\n",
       "2020-07-27/2020-08-02    1261.0\n",
       "2020-08-03/2020-08-09    1262.0\n",
       "Freq: W-SUN, Name: AI, dtype: float64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plot_df1['AI'].cumsum()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.6",
   "language": "python",
   "name": "py36"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  },
  "toc-autonumbering": true
 },
 "nbformat": 4,
 "nbformat_minor": 4
}