aux_ai_index_data.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# AI index data\n",
    "\n",
    "We generate regional activity statistics for the AI index"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preamble"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import datetime\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "today = str(datetime.date.today())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Functions\n",
    "\n",
    "def get_cited_papers(data,citation_var,q=0.75):\n",
    "    '''\n",
    "    This function subsets a dataset returning the most cited papers of the period (based on the citation variable and the quantile)\n",
    "    \n",
    "    '''\n",
    "    #Get the quantile\n",
    "    citation_quantile = data[citation_var].quantile(q)\n",
    "    \n",
    "    #Subset the data\n",
    "    subset_data = data.loc[data[citation_var]>=citation_quantile]\n",
    "\n",
    "    return(subset_data)\n",
    "\n",
    "\n",
    "def get_high_cited_year(data,q,year_lims):\n",
    "    '''\n",
    "    This function extracts high cited papers by year (to control for citation times).\n",
    "    \n",
    "    '''\n",
    "    #This loops over the years and extracts papers in the top quantile of activity.\n",
    "    \n",
    "    out = pd.concat([get_cited_papers(data.loc[data['year_created']==x,:],'citation_count',\n",
    "                                      q) for x in np.arange(year_lims[0],year_lims[1])])\n",
    "    \n",
    "    return(out)\n",
    "\n",
    "def create_lq_df(df,year=None):\n",
    "    '''\n",
    "    Takes a df with cells = activity in col in row and returns a df with cells = lq\n",
    "    \n",
    "    '''\n",
    "    \n",
    "    area_activity = df.sum(axis=0)\n",
    "    area_shares = area_activity/area_activity.sum()\n",
    "    \n",
    "    lqs = df.apply(lambda x: (x/x.sum())/area_shares, axis=1)\n",
    "    \n",
    "    if year!=None:\n",
    "        lqs['period'] = year\n",
    "    \n",
    "    return(lqs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "arx_papers = pd.read_csv('input_data/23_7_2019_ai_index_papers.csv',compression='zip')"
   ]
  },
  {
   "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>Unnamed: 0</th>\n",
       "      <th>Unnamed: 0.1</th>\n",
       "      <th>article_id</th>\n",
       "      <th>article_created</th>\n",
       "      <th>article_updated</th>\n",
       "      <th>article_title</th>\n",
       "      <th>journal_ref</th>\n",
       "      <th>doi</th>\n",
       "      <th>summary</th>\n",
       "      <th>citation_count</th>\n",
       "      <th>...</th>\n",
       "      <th>institute_name</th>\n",
       "      <th>institute_city</th>\n",
       "      <th>institute_state</th>\n",
       "      <th>institute_lat</th>\n",
       "      <th>institute_lon</th>\n",
       "      <th>year_created</th>\n",
       "      <th>institute_country</th>\n",
       "      <th>institute_region</th>\n",
       "      <th>country_reg</th>\n",
       "      <th>dl_cat</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0704.0002</td>\n",
       "      <td>2007-03-30</td>\n",
       "      <td>2008-12-13</td>\n",
       "      <td>Sparsity-certifying Graph Decompositions</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>We describe a new algorithm, the $(k,\\ell)$-...</td>\n",
       "      <td>18.0</td>\n",
       "      <td>...</td>\n",
       "      <td>Smith College</td>\n",
       "      <td>Northampton</td>\n",
       "      <td>Massachusetts</td>\n",
       "      <td>42.318119</td>\n",
       "      <td>-72.638139</td>\n",
       "      <td>2007</td>\n",
       "      <td>United States of America</td>\n",
       "      <td>Massachusetts</td>\n",
       "      <td>US_Massachusetts</td>\n",
       "      <td>not_dl</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0705.1541</td>\n",
       "      <td>2007-05-10</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unfolding Manhattan Towers</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>We provide an algorithm for unfolding the su...</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>Smith College</td>\n",
       "      <td>Northampton</td>\n",
       "      <td>Massachusetts</td>\n",
       "      <td>42.318119</td>\n",
       "      <td>-72.638139</td>\n",
       "      <td>2007</td>\n",
       "      <td>United States of America</td>\n",
       "      <td>Massachusetts</td>\n",
       "      <td>US_Massachusetts</td>\n",
       "      <td>not_dl</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0707.4258</td>\n",
       "      <td>2007-07-28</td>\n",
       "      <td>2009-06-24</td>\n",
       "      <td>Star Unfolding Convex Polyhedra via Quasigeode...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>We extend the notion of star unfolding to be...</td>\n",
       "      <td>12.0</td>\n",
       "      <td>...</td>\n",
       "      <td>Smith College</td>\n",
       "      <td>Northampton</td>\n",
       "      <td>Massachusetts</td>\n",
       "      <td>42.318119</td>\n",
       "      <td>-72.638139</td>\n",
       "      <td>2007</td>\n",
       "      <td>United States of America</td>\n",
       "      <td>Massachusetts</td>\n",
       "      <td>US_Massachusetts</td>\n",
       "      <td>not_dl</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>0709.1942</td>\n",
       "      <td>2007-09-12</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Connecting Polygonizations via Stretches and T...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>We show that the space of polygonizations of...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>Smith College</td>\n",
       "      <td>Northampton</td>\n",
       "      <td>Massachusetts</td>\n",
       "      <td>42.318119</td>\n",
       "      <td>-72.638139</td>\n",
       "      <td>2007</td>\n",
       "      <td>United States of America</td>\n",
       "      <td>Massachusetts</td>\n",
       "      <td>US_Massachusetts</td>\n",
       "      <td>not_dl</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>0709.3554</td>\n",
       "      <td>2007-09-21</td>\n",
       "      <td>NaN</td>\n",
       "      <td>A New Lower Bound on Guard Placement for Wirel...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>The problem of wireless localization asks to...</td>\n",
       "      <td>8.0</td>\n",
       "      <td>...</td>\n",
       "      <td>Smith College</td>\n",
       "      <td>Northampton</td>\n",
       "      <td>Massachusetts</td>\n",
       "      <td>42.318119</td>\n",
       "      <td>-72.638139</td>\n",
       "      <td>2007</td>\n",
       "      <td>United States of America</td>\n",
       "      <td>Massachusetts</td>\n",
       "      <td>US_Massachusetts</td>\n",
       "      <td>not_dl</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 25 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0  Unnamed: 0.1 article_id article_created article_updated  \\\n",
       "0           0             0  0704.0002      2007-03-30      2008-12-13   \n",
       "1           1             1  0705.1541      2007-05-10             NaN   \n",
       "2           2             2  0707.4258      2007-07-28      2009-06-24   \n",
       "3           3             3  0709.1942      2007-09-12             NaN   \n",
       "4           4             4  0709.3554      2007-09-21             NaN   \n",
       "\n",
       "                                       article_title journal_ref  doi  \\\n",
       "0           Sparsity-certifying Graph Decompositions         NaN  NaN   \n",
       "1                         Unfolding Manhattan Towers         NaN  NaN   \n",
       "2  Star Unfolding Convex Polyhedra via Quasigeode...         NaN  NaN   \n",
       "3  Connecting Polygonizations via Stretches and T...         NaN  NaN   \n",
       "4  A New Lower Bound on Guard Placement for Wirel...         NaN  NaN   \n",
       "\n",
       "                                             summary  citation_count  ...  \\\n",
       "0    We describe a new algorithm, the $(k,\\ell)$-...            18.0  ...   \n",
       "1    We provide an algorithm for unfolding the su...            15.0  ...   \n",
       "2    We extend the notion of star unfolding to be...            12.0  ...   \n",
       "3    We show that the space of polygonizations of...             4.0  ...   \n",
       "4    The problem of wireless localization asks to...             8.0  ...   \n",
       "\n",
       "  institute_name institute_city  institute_state  institute_lat institute_lon  \\\n",
       "0  Smith College    Northampton    Massachusetts      42.318119    -72.638139   \n",
       "1  Smith College    Northampton    Massachusetts      42.318119    -72.638139   \n",
       "2  Smith College    Northampton    Massachusetts      42.318119    -72.638139   \n",
       "3  Smith College    Northampton    Massachusetts      42.318119    -72.638139   \n",
       "4  Smith College    Northampton    Massachusetts      42.318119    -72.638139   \n",
       "\n",
       "  year_created         institute_country institute_region       country_reg  \\\n",
       "0         2007  United States of America    Massachusetts  US_Massachusetts   \n",
       "1         2007  United States of America    Massachusetts  US_Massachusetts   \n",
       "2         2007  United States of America    Massachusetts  US_Massachusetts   \n",
       "3         2007  United States of America    Massachusetts  US_Massachusetts   \n",
       "4         2007  United States of America    Massachusetts  US_Massachusetts   \n",
       "\n",
       "   dl_cat  \n",
       "0  not_dl  \n",
       "1  not_dl  \n",
       "2  not_dl  \n",
       "3  not_dl  \n",
       "4  not_dl  \n",
       "\n",
       "[5 rows x 25 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "arx_papers.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This data has already been labelled with DL / non DL categories. We use it to produce regional counts and RCA indices"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Set some parameters for analysis\n",
    "\n",
    "These could be changed to implement a country-level analysis, create different citation thresholds and thresholds for discretisation etc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Geography to focus on\n",
    "geo = 'institute_region'\n",
    "\n",
    "#You could also do 'institute_country'\n",
    "\n",
    "#Rank in the year citation\n",
    "high_cited = 0.5\n",
    "\n",
    "#Number of locations to focus on (ranked by activity)\n",
    "top_l = 40\n",
    "\n",
    "#Years to focus on\n",
    "focus_years = [2000,2019]\n",
    "\n",
    "#Threshold for discretising years\n",
    "year_thres = 2012\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Additional processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "California       11361\n",
       "Massachusetts     5424\n",
       "Beijing           4365\n",
       "New York          4124\n",
       "Ontario           3532\n",
       "Name: institute_region, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#List of high activity regions (we use it to subset the RCAs and focus on high activity locations)\n",
    "locs = arx_papers[geo].value_counts()\n",
    "locs.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "top_locs = locs.index[:top_l]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# We focus on high cited (above median) papers to capture influence / quality rather than quantity\n",
    "arx_high = get_high_cited_year(arx_papers,q=high_cited,year_lims=focus_years)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Yearly"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Paper counts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "high_counts = pd.concat([pd.crosstab(arx_high.loc[arx_high['year_created']==y,geo],\n",
    "                           arx_high.loc[arx_high['year_created']==y,'dl_cat'])['dl'] for y in np.arange(2008,2019)],axis=1,sort=True).fillna(0)\n",
    "\n",
    "high_counts.columns = np.arange(2008,2019)"
   ]
  },
  {
   "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>2008</th>\n",
       "      <th>2009</th>\n",
       "      <th>2010</th>\n",
       "      <th>2011</th>\n",
       "      <th>2012</th>\n",
       "      <th>2013</th>\n",
       "      <th>2014</th>\n",
       "      <th>2015</th>\n",
       "      <th>2016</th>\n",
       "      <th>2017</th>\n",
       "      <th>2018</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>California</th>\n",
       "      <td>3.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>203.0</td>\n",
       "      <td>335.0</td>\n",
       "      <td>563.0</td>\n",
       "      <td>717.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Beijing</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>54.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>189.0</td>\n",
       "      <td>336.0</td>\n",
       "      <td>515.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Massachusetts</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>92.0</td>\n",
       "      <td>127.0</td>\n",
       "      <td>195.0</td>\n",
       "      <td>322.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>New York</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>98.0</td>\n",
       "      <td>155.0</td>\n",
       "      <td>243.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Maryland</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>47.0</td>\n",
       "      <td>84.0</td>\n",
       "      <td>139.0</td>\n",
       "      <td>202.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>South West Community Development Council</th>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>61.0</td>\n",
       "      <td>86.0</td>\n",
       "      <td>112.0</td>\n",
       "      <td>195.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Guangdong</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>43.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>186.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Illinois</th>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>47.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>162.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Texas</th>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>34.0</td>\n",
       "      <td>39.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>148.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Shanghai</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>97.0</td>\n",
       "      <td>140.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                          2008  2009  2010  2011  2012  2013  \\\n",
       "California                                 3.0   8.0  24.0  31.0  32.0  36.0   \n",
       "Beijing                                    0.0   1.0   5.0   2.0  10.0  12.0   \n",
       "Massachusetts                              1.0   4.0   3.0  10.0  27.0  19.0   \n",
       "New York                                   1.0   2.0   2.0  11.0   8.0  28.0   \n",
       "Maryland                                   0.0   0.0   3.0   1.0  10.0  15.0   \n",
       "South West Community Development Council   0.0   2.0   2.0   3.0  17.0  16.0   \n",
       "Guangdong                                  0.0   0.0   0.0   0.0   4.0   0.0   \n",
       "Illinois                                   0.0   2.0   2.0   6.0  15.0  12.0   \n",
       "Texas                                      3.0   6.0   2.0   5.0  10.0   8.0   \n",
       "Shanghai                                   0.0   1.0   1.0   0.0   3.0   5.0   \n",
       "\n",
       "                                           2014   2015   2016   2017   2018  \n",
       "California                                100.0  203.0  335.0  563.0  717.0  \n",
       "Beijing                                    54.0   90.0  189.0  336.0  515.0  \n",
       "Massachusetts                              37.0   92.0  127.0  195.0  322.0  \n",
       "New York                                   29.0   64.0   98.0  155.0  243.0  \n",
       "Maryland                                   22.0   47.0   84.0  139.0  202.0  \n",
       "South West Community Development Council   27.0   61.0   86.0  112.0  195.0  \n",
       "Guangdong                                   4.0   35.0   43.0   74.0  186.0  \n",
       "Illinois                                   11.0   47.0   67.0   96.0  162.0  \n",
       "Texas                                      34.0   39.0   45.0   96.0  148.0  \n",
       "Shanghai                                   13.0   15.0   45.0   97.0  140.0  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "counts_year = high_counts.loc[high_counts.sum(axis=1)>0].sort_values(2018,ascending=False)\n",
    "\n",
    "counts_year.head(n=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x10dbbc668>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "counts_year.T.plot(legend=False,cmap='seismic_r',title='Paper activity per year',alpha=0.7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "counts_year.to_csv(f'output_data/{today}_counts_year.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## RCAs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "rca_year = pd.concat([create_lq_df(pd.crosstab(arx_high.loc[arx_high['year_created']==y,geo],\n",
    "                           arx_high.loc[arx_high['year_created']==y,'dl_cat']))['dl'] for y in np.arange(2008,2019)],axis=1,sort=True).fillna(0)\n",
    "\n",
    "rca_year.columns = np.arange(2008,2019)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>2008</th>\n",
       "      <th>2009</th>\n",
       "      <th>2010</th>\n",
       "      <th>2011</th>\n",
       "      <th>2012</th>\n",
       "      <th>2013</th>\n",
       "      <th>2014</th>\n",
       "      <th>2015</th>\n",
       "      <th>2016</th>\n",
       "      <th>2017</th>\n",
       "      <th>2018</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Beijing</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.073803</td>\n",
       "      <td>2.167677</td>\n",
       "      <td>0.464449</td>\n",
       "      <td>1.341593</td>\n",
       "      <td>0.962417</td>\n",
       "      <td>1.872748</td>\n",
       "      <td>1.315873</td>\n",
       "      <td>1.590960</td>\n",
       "      <td>1.488500</td>\n",
       "      <td>1.522979</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Quebec</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.593176</td>\n",
       "      <td>2.491530</td>\n",
       "      <td>3.247430</td>\n",
       "      <td>2.683011</td>\n",
       "      <td>2.247985</td>\n",
       "      <td>1.731164</td>\n",
       "      <td>1.521855</td>\n",
       "      <td>1.379772</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Australian Capital Territory</th>\n",
       "      <td>5.191824</td>\n",
       "      <td>5.624680</td>\n",
       "      <td>2.341091</td>\n",
       "      <td>2.593176</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.007462</td>\n",
       "      <td>1.031126</td>\n",
       "      <td>1.517213</td>\n",
       "      <td>1.187902</td>\n",
       "      <td>1.030071</td>\n",
       "      <td>1.328813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>South West Community Development Council</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.629218</td>\n",
       "      <td>0.907400</td>\n",
       "      <td>0.686429</td>\n",
       "      <td>1.546915</td>\n",
       "      <td>1.334145</td>\n",
       "      <td>1.202981</td>\n",
       "      <td>1.412528</td>\n",
       "      <td>1.330650</td>\n",
       "      <td>1.145302</td>\n",
       "      <td>1.293290</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Royal Borough of Kensington and Chelsea</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.196850</td>\n",
       "      <td>2.790514</td>\n",
       "      <td>0.656649</td>\n",
       "      <td>0.763797</td>\n",
       "      <td>0.960184</td>\n",
       "      <td>1.162894</td>\n",
       "      <td>1.154598</td>\n",
       "      <td>1.272270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Maryland</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.950909</td>\n",
       "      <td>0.471487</td>\n",
       "      <td>1.835864</td>\n",
       "      <td>2.020460</td>\n",
       "      <td>1.833113</td>\n",
       "      <td>1.623735</td>\n",
       "      <td>1.524793</td>\n",
       "      <td>1.415805</td>\n",
       "      <td>1.265970</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Shanghai</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.624851</td>\n",
       "      <td>1.026794</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.255731</td>\n",
       "      <td>1.459221</td>\n",
       "      <td>1.737639</td>\n",
       "      <td>0.905048</td>\n",
       "      <td>1.419580</td>\n",
       "      <td>1.476200</td>\n",
       "      <td>1.250159</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cambridgeshire</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.481815</td>\n",
       "      <td>2.092885</td>\n",
       "      <td>0.700426</td>\n",
       "      <td>1.573710</td>\n",
       "      <td>1.427965</td>\n",
       "      <td>1.066922</td>\n",
       "      <td>1.231978</td>\n",
       "      <td>1.193302</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>South Australia</th>\n",
       "      <td>2.396226</td>\n",
       "      <td>5.062212</td>\n",
       "      <td>2.601212</td>\n",
       "      <td>1.637795</td>\n",
       "      <td>1.006195</td>\n",
       "      <td>2.529316</td>\n",
       "      <td>2.105216</td>\n",
       "      <td>1.427965</td>\n",
       "      <td>1.317268</td>\n",
       "      <td>1.347157</td>\n",
       "      <td>1.189744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>California</th>\n",
       "      <td>0.543330</td>\n",
       "      <td>0.787455</td>\n",
       "      <td>1.463182</td>\n",
       "      <td>1.422804</td>\n",
       "      <td>0.664408</td>\n",
       "      <td>0.642156</td>\n",
       "      <td>1.156017</td>\n",
       "      <td>1.223110</td>\n",
       "      <td>1.183155</td>\n",
       "      <td>1.162464</td>\n",
       "      <td>1.178109</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              2008      2009      2010  \\\n",
       "Beijing                                   0.000000  1.073803  2.167677   \n",
       "Quebec                                    0.000000  0.000000  0.000000   \n",
       "Australian Capital Territory              5.191824  5.624680  2.341091   \n",
       "South West Community Development Council  0.000000  1.629218  0.907400   \n",
       "Royal Borough of Kensington and Chelsea   0.000000  0.000000  0.000000   \n",
       "Maryland                                  0.000000  0.000000  1.950909   \n",
       "Shanghai                                  0.000000  2.624851  1.026794   \n",
       "Cambridgeshire                            0.000000  0.000000  0.000000   \n",
       "South Australia                           2.396226  5.062212  2.601212   \n",
       "California                                0.543330  0.787455  1.463182   \n",
       "\n",
       "                                              2011      2012      2013  \\\n",
       "Beijing                                   0.464449  1.341593  0.962417   \n",
       "Quebec                                    2.593176  2.491530  3.247430   \n",
       "Australian Capital Territory              2.593176  0.000000  1.007462   \n",
       "South West Community Development Council  0.686429  1.546915  1.334145   \n",
       "Royal Borough of Kensington and Chelsea   1.196850  2.790514  0.656649   \n",
       "Maryland                                  0.471487  1.835864  2.020460   \n",
       "Shanghai                                  0.000000  1.255731  1.459221   \n",
       "Cambridgeshire                            1.481815  2.092885  0.700426   \n",
       "South Australia                           1.637795  1.006195  2.529316   \n",
       "California                                1.422804  0.664408  0.642156   \n",
       "\n",
       "                                              2014      2015      2016  \\\n",
       "Beijing                                   1.872748  1.315873  1.590960   \n",
       "Quebec                                    2.683011  2.247985  1.731164   \n",
       "Australian Capital Territory              1.031126  1.517213  1.187902   \n",
       "South West Community Development Council  1.202981  1.412528  1.330650   \n",
       "Royal Borough of Kensington and Chelsea   0.763797  0.960184  1.162894   \n",
       "Maryland                                  1.833113  1.623735  1.524793   \n",
       "Shanghai                                  1.737639  0.905048  1.419580   \n",
       "Cambridgeshire                            1.573710  1.427965  1.066922   \n",
       "South Australia                           2.105216  1.427965  1.317268   \n",
       "California                                1.156017  1.223110  1.183155   \n",
       "\n",
       "                                              2017      2018  \n",
       "Beijing                                   1.488500  1.522979  \n",
       "Quebec                                    1.521855  1.379772  \n",
       "Australian Capital Territory              1.030071  1.328813  \n",
       "South West Community Development Council  1.145302  1.293290  \n",
       "Royal Borough of Kensington and Chelsea   1.154598  1.272270  \n",
       "Maryland                                  1.415805  1.265970  \n",
       "Shanghai                                  1.476200  1.250159  \n",
       "Cambridgeshire                            1.231978  1.193302  \n",
       "South Australia                           1.347157  1.189744  \n",
       "California                                1.162464  1.178109  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#LQs focusing on the top 100 regions\n",
    "rca_year.loc[top_locs].sort_values(2018,ascending=False).head(n=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is likely to be noisy on a yearly basis so perhaps better to smooth it.\n",
    "\n",
    "Also produce the discretised table with before / after 2012"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "rca_year.to_csv(f'output_data/{today}_rca_year.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pre / post 2012"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Add a variable for discretising\n",
    "arx_high['discr'] = [f'pre_{year_thres}' if x<=2012 else f'post_{year_thres}' for x in arx_high['year_created']]\n",
    "\n",
    "discr_vars = [f'pre_{year_thres}',f'post_{year_thres}']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Counts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "counts_discr = pd.concat([pd.crosstab(arx_high.loc[arx_high['discr']==discr,geo],\n",
    "                           arx_high.loc[arx_high['discr']==discr,'dl_cat'])['dl'] for discr in discr_vars],axis=1,sort=True).fillna(0)\n",
    "\n",
    "counts_discr.columns = discr_vars"
   ]
  },
  {
   "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>pre_2012</th>\n",
       "      <th>post_2012</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>California</th>\n",
       "      <td>105.0</td>\n",
       "      <td>1954.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Beijing</th>\n",
       "      <td>20.0</td>\n",
       "      <td>1196.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Massachusetts</th>\n",
       "      <td>47.0</td>\n",
       "      <td>792.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>New York</th>\n",
       "      <td>28.0</td>\n",
       "      <td>617.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Maryland</th>\n",
       "      <td>20.0</td>\n",
       "      <td>509.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>South West Community Development Council</th>\n",
       "      <td>24.0</td>\n",
       "      <td>497.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Illinois</th>\n",
       "      <td>28.0</td>\n",
       "      <td>395.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bavaria</th>\n",
       "      <td>16.0</td>\n",
       "      <td>384.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Texas</th>\n",
       "      <td>28.0</td>\n",
       "      <td>370.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ontario</th>\n",
       "      <td>22.0</td>\n",
       "      <td>362.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                          pre_2012  post_2012\n",
       "California                                   105.0     1954.0\n",
       "Beijing                                       20.0     1196.0\n",
       "Massachusetts                                 47.0      792.0\n",
       "New York                                      28.0      617.0\n",
       "Maryland                                      20.0      509.0\n",
       "South West Community Development Council      24.0      497.0\n",
       "Illinois                                      28.0      395.0\n",
       "Bavaria                                       16.0      384.0\n",
       "Texas                                         28.0      370.0\n",
       "Ontario                                       22.0      362.0"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "counts_discr.sort_values(discr_vars[1],ascending=False).head(n=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "counts_discr.to_csv(f'output_data/{today}_counts_discretised.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LQ"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "lq_discr = pd.concat([create_lq_df(pd.crosstab(arx_high.loc[arx_high['discr']==discr,geo],\n",
    "                           arx_high.loc[arx_high['discr']==discr,'dl_cat']))['dl'] for discr in discr_vars],axis=1,sort=True).fillna(0)\n",
    "\n",
    "lq_discr.columns = discr_vars"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>pre_2012</th>\n",
       "      <th>post_2012</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Quebec</th>\n",
       "      <td>1.923416</td>\n",
       "      <td>1.684241</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Beijing</th>\n",
       "      <td>1.313052</td>\n",
       "      <td>1.588248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Maryland</th>\n",
       "      <td>1.649563</td>\n",
       "      <td>1.473750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Shanghai</th>\n",
       "      <td>0.976973</td>\n",
       "      <td>1.426469</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>South Australia</th>\n",
       "      <td>1.823683</td>\n",
       "      <td>1.368928</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>South West Community Development Council</th>\n",
       "      <td>1.283113</td>\n",
       "      <td>1.242089</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Oxfordshire</th>\n",
       "      <td>0.693513</td>\n",
       "      <td>1.228139</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Royal Borough of Kensington and Chelsea</th>\n",
       "      <td>1.414927</td>\n",
       "      <td>1.213193</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cambridgeshire</th>\n",
       "      <td>1.504539</td>\n",
       "      <td>1.185639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Canton of ZÃ¼rich</th>\n",
       "      <td>0.923684</td>\n",
       "      <td>1.154979</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                          pre_2012  post_2012\n",
       "Quebec                                    1.923416   1.684241\n",
       "Beijing                                   1.313052   1.588248\n",
       "Maryland                                  1.649563   1.473750\n",
       "Shanghai                                  0.976973   1.426469\n",
       "South Australia                           1.823683   1.368928\n",
       "South West Community Development Council  1.283113   1.242089\n",
       "Oxfordshire                               0.693513   1.228139\n",
       "Royal Borough of Kensington and Chelsea   1.414927   1.213193\n",
       "Cambridgeshire                            1.504539   1.185639\n",
       "Canton of ZÃ¼rich                         0.923684   1.154979"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lq_discr.loc[top_locs].sort_values(discr_vars[1],ascending=False).head(n=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "lq_discr.to_csv(f'output_data/{today}_lq_discretised.csv')"
   ]
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}