standard_datasets_results.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Results of indexing tests on standard datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import datetime, pathlib, sys\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from IPython import display\n",
    "import ipywidgets as ipw\n",
    "\n",
    "sys.path.insert(0, '../../src')\n",
    "from indexing import ExtConvexHull\n",
    "\n",
    "sns.set_style()\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Helpers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plot_dataset_index_stats(indexing_stats):\n",
    "    markersize = 10\n",
    "    plot_kwargs = {'FullIndex': {'label': 'Full', 'marker': '+', 'markersize':markersize},\n",
    "#                   'AdjacentBlockIndex': {'label':'Neighbourhood Blocking', 'marker': 'o', 'markersize':markersize},\n",
    "                   'NeighbourhoodBlockIndex': {'label':'Neighbourhood Blocking', 'marker': 'o', 'markersize':markersize},\n",
    "                   'BlockIndex': {'label': 'Standard Blocking', 'marker': 's', 'markersize':markersize},\n",
    "                   'SortedNeighbourhoodIndex': {'label': 'Sorted Neighbourhood', 'marker': '^', 'markersize':markersize},\n",
    "                  }\n",
    "    merged_indexer_classes = {'NeighbourhoodBlockIndex': ['BlockIndex']}\n",
    "    indexing_stats = indexing_stats.copy()\n",
    "    indexing_stats = indexing_stats[~indexing_stats['failed']]\n",
    "#    indexing_stats['renamed_indexer_class'] = indexing_stats['indexer_class']\n",
    "#    indexing_stats['renamed_indexer_class'][indexing_stats['renamed_indexer_class']=='AdjacentBlockIndex'] = 'NeighbourhoodBlockIndex'\n",
    "#    indexing_stats = indexing_stats[indexing_stats['max_non_matches'].isnull() | (indexing_stats['max_non_matches'] == 0)]\n",
    "#    indexing_stats = indexing_stats[indexing_stats['on'].map(lambda x:(x is np.nan) or 'soc_sec_id' not in x)]\n",
    "    axis_vars = ['reduction_ratio', 'recall']\n",
    "    plt.figure(figsize=(7,7))\n",
    "    ax = plt.gca()\n",
    "    for axis_name, axis_var in zip(['x', 'y'], axis_vars):\n",
    "        getattr(ax, 'set_{axis_name}label'.format(**locals()))('{} (truncated)'.format(axis_var.replace('_', ' ').title()))\n",
    "        getattr(ax, 'set_{axis_name}lim'.format(**locals()))(0.5, 1.05)\n",
    "    for indexer_class, vals in indexing_stats.groupby('indexer_class'):\n",
    "        vals = pd.concat([vals] + [indexing_stats[indexing_stats['indexer_class'] == cls] for cls in merged_indexer_classes.get(indexer_class, [])])\n",
    "        cost_hull = ExtConvexHull(points = (1-vals[axis_vars]).values)\n",
    "        frontier_points = pd.DataFrame((1-cost_hull.points[cost_hull.minimizing_points()]), columns=['x', 'y']).sort_values('x')\n",
    "        ax.plot(*frontier_points.values.T, color='grey', **plot_kwargs[indexer_class])\n",
    "#        ax.scatter(*[vals[col].values for col in axis_vars], color='grey', **plot_kwargs[indexer_class])\n",
    "    plt.legend(loc='lower left', title='Index Type')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6b161d46c8874549a867ee7a7d885896"
      }
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def get_indexing_stats(database_key, cache_dir='.'):\n",
    "    source_file = pathlib.Path(cache_dir) / 'indexing_stats_{database_key}.pickle'.format(**locals())\n",
    "    return pd.read_pickle(str(source_file))\n",
    "\n",
    "def get_database_keys(cache_dir='.'):\n",
    "    prefix = 'indexing_stats_'\n",
    "    suffix = '.pickle'\n",
    "    return sorted(p.name.replace(prefix,'').replace(suffix, '') for p in pathlib.Path(cache_dir).glob('{prefix}*{suffix}'.format(prefix=prefix, suffix=suffix)))\n",
    "\n",
    "cache_dir = '.'\n",
    "#cache_dir = 'indexing_stats_imported'\n",
    "\n",
    "def do_index_stats_plot(database_key):\n",
    "    if database_key:\n",
    "        plot_dataset_index_stats(get_indexing_stats(database_key, cache_dir=cache_dir))\n",
    "    \n",
    "ipw.interact(do_index_stats_plot, database_key = [None] + get_database_keys(cache_dir=cache_dir\n",
    "                                                                           ));"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "214\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/dan/anaconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: 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: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>dataset</th>\n",
       "      <th>description</th>\n",
       "      <th>elapsed_seconds</th>\n",
       "      <th>failed</th>\n",
       "      <th>index_size</th>\n",
       "      <th>indexer_class</th>\n",
       "      <th>max_non_matches</th>\n",
       "      <th>max_nulls</th>\n",
       "      <th>max_rank_differences</th>\n",
       "      <th>on</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "      <th>reduction_ratio</th>\n",
       "      <th>window</th>\n",
       "      <th>renamed_indexer_class</th>\n",
       "      <th>recall+reduction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>483</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['soc_sec_id'])</td>\n",
       "      <td>0.069761</td>\n",
       "      <td>False</td>\n",
       "      <td>1686.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[soc_sec_id]</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.871768</td>\n",
       "      <td>0.999865</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.871633</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>189</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day', 'date_of_bir...</td>\n",
       "      <td>0.295529</td>\n",
       "      <td>False</td>\n",
       "      <td>5647.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day, date_of_birth_year]</td>\n",
       "      <td>0.297149</td>\n",
       "      <td>0.867632</td>\n",
       "      <td>0.999548</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.867180</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>204</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_month', 'date_of_b...</td>\n",
       "      <td>0.337197</td>\n",
       "      <td>False</td>\n",
       "      <td>11812.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_month, date_of_birth_year]</td>\n",
       "      <td>0.142059</td>\n",
       "      <td>0.867632</td>\n",
       "      <td>0.999055</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.866687</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>128</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day', 'date_of_bir...</td>\n",
       "      <td>0.395938</td>\n",
       "      <td>False</td>\n",
       "      <td>34249.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day, date_of_birth_month]</td>\n",
       "      <td>0.048907</td>\n",
       "      <td>0.866081</td>\n",
       "      <td>0.997260</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.863340</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>187</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day', 'date_of_bir...</td>\n",
       "      <td>0.320731</td>\n",
       "      <td>False</td>\n",
       "      <td>1979.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day, date_of_birth_orig]</td>\n",
       "      <td>0.843355</td>\n",
       "      <td>0.862978</td>\n",
       "      <td>0.999842</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.862820</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>115</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day', 'date_of_bir...</td>\n",
       "      <td>0.317954</td>\n",
       "      <td>False</td>\n",
       "      <td>1979.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day, date_of_birth_month, date_...</td>\n",
       "      <td>0.843355</td>\n",
       "      <td>0.862978</td>\n",
       "      <td>0.999842</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.862820</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>127</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_month', 'date_of_b...</td>\n",
       "      <td>0.317666</td>\n",
       "      <td>False</td>\n",
       "      <td>1979.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_month, date_of_birth_orig, date...</td>\n",
       "      <td>0.843355</td>\n",
       "      <td>0.862978</td>\n",
       "      <td>0.999842</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.862820</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>140</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_month', 'date_of_b...</td>\n",
       "      <td>0.331411</td>\n",
       "      <td>False</td>\n",
       "      <td>1979.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_month, date_of_birth_orig]</td>\n",
       "      <td>0.843355</td>\n",
       "      <td>0.862978</td>\n",
       "      <td>0.999842</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.862820</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>220</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_orig', 'date_of_bi...</td>\n",
       "      <td>0.309447</td>\n",
       "      <td>False</td>\n",
       "      <td>1979.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_orig, date_of_birth_year]</td>\n",
       "      <td>0.843355</td>\n",
       "      <td>0.862978</td>\n",
       "      <td>0.999842</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.862820</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>482</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day'])</td>\n",
       "      <td>1.800290</td>\n",
       "      <td>False</td>\n",
       "      <td>389619.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day]</td>\n",
       "      <td>0.004335</td>\n",
       "      <td>0.873320</td>\n",
       "      <td>0.968824</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.842144</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>239</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_year', 'state'])</td>\n",
       "      <td>0.370920</td>\n",
       "      <td>False</td>\n",
       "      <td>28400.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_year, state]</td>\n",
       "      <td>0.056056</td>\n",
       "      <td>0.823164</td>\n",
       "      <td>0.997728</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.820892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>268</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['soc_sec_id', 'state'])</td>\n",
       "      <td>0.306980</td>\n",
       "      <td>False</td>\n",
       "      <td>1580.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[soc_sec_id, state]</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.816960</td>\n",
       "      <td>0.999874</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.816833</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>196</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day', 'state'])</td>\n",
       "      <td>0.508658</td>\n",
       "      <td>False</td>\n",
       "      <td>88351.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day, state]</td>\n",
       "      <td>0.018019</td>\n",
       "      <td>0.823164</td>\n",
       "      <td>0.992931</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.816095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>227</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_orig', 'state'])</td>\n",
       "      <td>0.296237</td>\n",
       "      <td>False</td>\n",
       "      <td>1643.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_orig, state]</td>\n",
       "      <td>0.957395</td>\n",
       "      <td>0.813340</td>\n",
       "      <td>0.999869</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.813209</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>212</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_month', 'state'])</td>\n",
       "      <td>1.211830</td>\n",
       "      <td>False</td>\n",
       "      <td>222207.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_month, state]</td>\n",
       "      <td>0.007160</td>\n",
       "      <td>0.822647</td>\n",
       "      <td>0.982220</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.804867</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>194</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day', 'soc_sec_id'])</td>\n",
       "      <td>0.294836</td>\n",
       "      <td>False</td>\n",
       "      <td>1472.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day, soc_sec_id]</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.761117</td>\n",
       "      <td>0.999882</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.760999</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>489</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['postcode'])</td>\n",
       "      <td>0.102188</td>\n",
       "      <td>False</td>\n",
       "      <td>14052.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[postcode]</td>\n",
       "      <td>0.104825</td>\n",
       "      <td>0.761634</td>\n",
       "      <td>0.998876</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.760510</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>138</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_year', 'soc_sec_id'])</td>\n",
       "      <td>0.325827</td>\n",
       "      <td>False</td>\n",
       "      <td>1471.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_year, soc_sec_id]</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.760600</td>\n",
       "      <td>0.999882</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.760482</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>210</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_month', 'soc_sec_i...</td>\n",
       "      <td>0.312079</td>\n",
       "      <td>False</td>\n",
       "      <td>1466.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_month, soc_sec_id]</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.758014</td>\n",
       "      <td>0.999883</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.757897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_month', 'date_of_b...</td>\n",
       "      <td>0.319502</td>\n",
       "      <td>False</td>\n",
       "      <td>1461.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_month, date_of_birth_year, soc_...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.755429</td>\n",
       "      <td>0.999883</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.755312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>225</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_orig', 'soc_sec_id'])</td>\n",
       "      <td>0.298602</td>\n",
       "      <td>False</td>\n",
       "      <td>1456.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_orig, soc_sec_id]</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.752844</td>\n",
       "      <td>0.999883</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.752727</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>486</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['street_number'])</td>\n",
       "      <td>0.784496</td>\n",
       "      <td>False</td>\n",
       "      <td>149898.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[street_number]</td>\n",
       "      <td>0.009687</td>\n",
       "      <td>0.750776</td>\n",
       "      <td>0.988006</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.738781</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>260</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['postcode', 'state'])</td>\n",
       "      <td>0.297163</td>\n",
       "      <td>False</td>\n",
       "      <td>4060.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[postcode, state]</td>\n",
       "      <td>0.338916</td>\n",
       "      <td>0.711479</td>\n",
       "      <td>0.999675</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.711154</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>276</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['state', 'street_number'])</td>\n",
       "      <td>0.379718</td>\n",
       "      <td>False</td>\n",
       "      <td>34151.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[state, street_number]</td>\n",
       "      <td>0.039882</td>\n",
       "      <td>0.704240</td>\n",
       "      <td>0.997267</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.701507</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>258</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['postcode', 'soc_sec_id'])</td>\n",
       "      <td>0.079076</td>\n",
       "      <td>False</td>\n",
       "      <td>1287.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[postcode, soc_sec_id]</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.665460</td>\n",
       "      <td>0.999897</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.665357</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>236</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_year', 'postcode'])</td>\n",
       "      <td>0.316014</td>\n",
       "      <td>False</td>\n",
       "      <td>1377.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_year, postcode]</td>\n",
       "      <td>0.930283</td>\n",
       "      <td>0.662358</td>\n",
       "      <td>0.999890</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.662248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>192</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day', 'postcode'])</td>\n",
       "      <td>0.312408</td>\n",
       "      <td>False</td>\n",
       "      <td>1658.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day, postcode]</td>\n",
       "      <td>0.772014</td>\n",
       "      <td>0.661841</td>\n",
       "      <td>0.999867</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.661708</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>208</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_month', 'postcode'])</td>\n",
       "      <td>0.330807</td>\n",
       "      <td>False</td>\n",
       "      <td>2230.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_month, postcode]</td>\n",
       "      <td>0.572646</td>\n",
       "      <td>0.660290</td>\n",
       "      <td>0.999822</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.660111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>223</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_orig', 'postcode'])</td>\n",
       "      <td>0.325566</td>\n",
       "      <td>False</td>\n",
       "      <td>1265.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_orig, postcode]</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.654085</td>\n",
       "      <td>0.999899</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.653984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_month', 'date_of_b...</td>\n",
       "      <td>0.319703</td>\n",
       "      <td>False</td>\n",
       "      <td>1265.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_month, date_of_birth_orig, date...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.654085</td>\n",
       "      <td>0.999899</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.653984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>270</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['soc_sec_id', 'street_number'])</td>\n",
       "      <td>0.324023</td>\n",
       "      <td>False</td>\n",
       "      <td>1258.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[soc_sec_id, street_number]</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.650465</td>\n",
       "      <td>0.999899</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.650365</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day', 'street_numb...</td>\n",
       "      <td>0.326778</td>\n",
       "      <td>False</td>\n",
       "      <td>5831.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day, street_number]</td>\n",
       "      <td>0.215229</td>\n",
       "      <td>0.648914</td>\n",
       "      <td>0.999533</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.648448</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>241</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_year', 'street_num...</td>\n",
       "      <td>0.325707</td>\n",
       "      <td>False</td>\n",
       "      <td>2686.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_year, street_number]</td>\n",
       "      <td>0.466121</td>\n",
       "      <td>0.647363</td>\n",
       "      <td>0.999785</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.647148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>214</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_month', 'street_nu...</td>\n",
       "      <td>0.373399</td>\n",
       "      <td>False</td>\n",
       "      <td>12905.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_month, street_number]</td>\n",
       "      <td>0.097094</td>\n",
       "      <td>0.647880</td>\n",
       "      <td>0.998967</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.646847</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>488</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['suburb'])</td>\n",
       "      <td>0.315807</td>\n",
       "      <td>False</td>\n",
       "      <td>14837.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[suburb]</td>\n",
       "      <td>0.084249</td>\n",
       "      <td>0.646329</td>\n",
       "      <td>0.998813</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.645142</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>229</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_orig', 'street_num...</td>\n",
       "      <td>0.297485</td>\n",
       "      <td>False</td>\n",
       "      <td>1245.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_orig, street_number]</td>\n",
       "      <td>0.992771</td>\n",
       "      <td>0.639090</td>\n",
       "      <td>0.999900</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.638990</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day', 'date_of_bir...</td>\n",
       "      <td>0.350730</td>\n",
       "      <td>False</td>\n",
       "      <td>1245.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day, date_of_birth_month, date_...</td>\n",
       "      <td>0.992771</td>\n",
       "      <td>0.639090</td>\n",
       "      <td>0.999900</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.638990</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>119</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day', 'state', 'st...</td>\n",
       "      <td>0.337206</td>\n",
       "      <td>False</td>\n",
       "      <td>2204.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day, state, street_number]</td>\n",
       "      <td>0.537205</td>\n",
       "      <td>0.612203</td>\n",
       "      <td>0.999824</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.612026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>135</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['state', 'suburb'])</td>\n",
       "      <td>0.327660</td>\n",
       "      <td>False</td>\n",
       "      <td>4117.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[state, suburb]</td>\n",
       "      <td>0.285159</td>\n",
       "      <td>0.607032</td>\n",
       "      <td>0.999671</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.606703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>485</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['given_name'])</td>\n",
       "      <td>0.415172</td>\n",
       "      <td>False</td>\n",
       "      <td>48883.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[given_name]</td>\n",
       "      <td>0.024139</td>\n",
       "      <td>0.610134</td>\n",
       "      <td>0.996089</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.606223</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>250</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['given_name', 'state'])</td>\n",
       "      <td>0.315181</td>\n",
       "      <td>False</td>\n",
       "      <td>11780.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[given_name, state]</td>\n",
       "      <td>0.093973</td>\n",
       "      <td>0.572389</td>\n",
       "      <td>0.999057</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.571446</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>272</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['soc_sec_id', 'suburb'])</td>\n",
       "      <td>0.301735</td>\n",
       "      <td>False</td>\n",
       "      <td>1089.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[soc_sec_id, suburb]</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.563082</td>\n",
       "      <td>0.999913</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.562995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>262</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['postcode', 'street_number'])</td>\n",
       "      <td>0.302725</td>\n",
       "      <td>False</td>\n",
       "      <td>1245.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[postcode, street_number]</td>\n",
       "      <td>0.869076</td>\n",
       "      <td>0.559462</td>\n",
       "      <td>0.999900</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.559363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>243</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_year', 'suburb'])</td>\n",
       "      <td>0.292492</td>\n",
       "      <td>False</td>\n",
       "      <td>1191.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_year, suburb]</td>\n",
       "      <td>0.904282</td>\n",
       "      <td>0.556877</td>\n",
       "      <td>0.999905</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.556782</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>216</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_month', 'suburb'])</td>\n",
       "      <td>0.301161</td>\n",
       "      <td>False</td>\n",
       "      <td>2108.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_month, suburb]</td>\n",
       "      <td>0.509962</td>\n",
       "      <td>0.555843</td>\n",
       "      <td>0.999831</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.555674</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>484</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['surname'])</td>\n",
       "      <td>0.440356</td>\n",
       "      <td>False</td>\n",
       "      <td>49781.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[surname]</td>\n",
       "      <td>0.021715</td>\n",
       "      <td>0.558945</td>\n",
       "      <td>0.996017</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.554962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_day', 'suburb'])</td>\n",
       "      <td>0.323354</td>\n",
       "      <td>False</td>\n",
       "      <td>1461.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_day, suburb]</td>\n",
       "      <td>0.733744</td>\n",
       "      <td>0.554292</td>\n",
       "      <td>0.999883</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.554175</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>231</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['date_of_birth_orig', 'suburb'])</td>\n",
       "      <td>0.300556</td>\n",
       "      <td>False</td>\n",
       "      <td>1063.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[date_of_birth_orig, suburb]</td>\n",
       "      <td>0.999059</td>\n",
       "      <td>0.549121</td>\n",
       "      <td>0.999915</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.549036</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>248</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['given_name', 'soc_sec_id'])</td>\n",
       "      <td>0.299519</td>\n",
       "      <td>False</td>\n",
       "      <td>1036.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[given_name, soc_sec_id]</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.535677</td>\n",
       "      <td>0.999917</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.535594</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>487</th>\n",
       "      <td>febrl2</td>\n",
       "      <td>BlockIndex (['address_1'])</td>\n",
       "      <td>0.312787</td>\n",
       "      <td>False</td>\n",
       "      <td>6148.0</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>[address_1]</td>\n",
       "      <td>0.167859</td>\n",
       "      <td>0.533609</td>\n",
       "      <td>0.999508</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BlockIndex</td>\n",
       "      <td>1.533117</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    dataset                                        description  \\\n",
       "483  febrl2                        BlockIndex (['soc_sec_id'])   \n",
       "189  febrl2  BlockIndex (['date_of_birth_day', 'date_of_bir...   \n",
       "204  febrl2  BlockIndex (['date_of_birth_month', 'date_of_b...   \n",
       "128  febrl2  BlockIndex (['date_of_birth_day', 'date_of_bir...   \n",
       "187  febrl2  BlockIndex (['date_of_birth_day', 'date_of_bir...   \n",
       "115  febrl2  BlockIndex (['date_of_birth_day', 'date_of_bir...   \n",
       "127  febrl2  BlockIndex (['date_of_birth_month', 'date_of_b...   \n",
       "140  febrl2  BlockIndex (['date_of_birth_month', 'date_of_b...   \n",
       "220  febrl2  BlockIndex (['date_of_birth_orig', 'date_of_bi...   \n",
       "482  febrl2                 BlockIndex (['date_of_birth_day'])   \n",
       "239  febrl2       BlockIndex (['date_of_birth_year', 'state'])   \n",
       "268  febrl2               BlockIndex (['soc_sec_id', 'state'])   \n",
       "196  febrl2        BlockIndex (['date_of_birth_day', 'state'])   \n",
       "227  febrl2       BlockIndex (['date_of_birth_orig', 'state'])   \n",
       "212  febrl2      BlockIndex (['date_of_birth_month', 'state'])   \n",
       "194  febrl2   BlockIndex (['date_of_birth_day', 'soc_sec_id'])   \n",
       "489  febrl2                          BlockIndex (['postcode'])   \n",
       "138  febrl2  BlockIndex (['date_of_birth_year', 'soc_sec_id'])   \n",
       "210  febrl2  BlockIndex (['date_of_birth_month', 'soc_sec_i...   \n",
       "125  febrl2  BlockIndex (['date_of_birth_month', 'date_of_b...   \n",
       "225  febrl2  BlockIndex (['date_of_birth_orig', 'soc_sec_id'])   \n",
       "486  febrl2                     BlockIndex (['street_number'])   \n",
       "260  febrl2                 BlockIndex (['postcode', 'state'])   \n",
       "276  febrl2            BlockIndex (['state', 'street_number'])   \n",
       "258  febrl2            BlockIndex (['postcode', 'soc_sec_id'])   \n",
       "236  febrl2    BlockIndex (['date_of_birth_year', 'postcode'])   \n",
       "192  febrl2     BlockIndex (['date_of_birth_day', 'postcode'])   \n",
       "208  febrl2   BlockIndex (['date_of_birth_month', 'postcode'])   \n",
       "223  febrl2    BlockIndex (['date_of_birth_orig', 'postcode'])   \n",
       "114  febrl2  BlockIndex (['date_of_birth_month', 'date_of_b...   \n",
       "270  febrl2       BlockIndex (['soc_sec_id', 'street_number'])   \n",
       "132  febrl2  BlockIndex (['date_of_birth_day', 'street_numb...   \n",
       "241  febrl2  BlockIndex (['date_of_birth_year', 'street_num...   \n",
       "214  febrl2  BlockIndex (['date_of_birth_month', 'street_nu...   \n",
       "488  febrl2                            BlockIndex (['suburb'])   \n",
       "229  febrl2  BlockIndex (['date_of_birth_orig', 'street_num...   \n",
       "102  febrl2  BlockIndex (['date_of_birth_day', 'date_of_bir...   \n",
       "119  febrl2  BlockIndex (['date_of_birth_day', 'state', 'st...   \n",
       "135  febrl2                   BlockIndex (['state', 'suburb'])   \n",
       "485  febrl2                        BlockIndex (['given_name'])   \n",
       "250  febrl2               BlockIndex (['given_name', 'state'])   \n",
       "272  febrl2              BlockIndex (['soc_sec_id', 'suburb'])   \n",
       "262  febrl2         BlockIndex (['postcode', 'street_number'])   \n",
       "243  febrl2      BlockIndex (['date_of_birth_year', 'suburb'])   \n",
       "216  febrl2     BlockIndex (['date_of_birth_month', 'suburb'])   \n",
       "484  febrl2                           BlockIndex (['surname'])   \n",
       "199  febrl2       BlockIndex (['date_of_birth_day', 'suburb'])   \n",
       "231  febrl2      BlockIndex (['date_of_birth_orig', 'suburb'])   \n",
       "248  febrl2          BlockIndex (['given_name', 'soc_sec_id'])   \n",
       "487  febrl2                         BlockIndex (['address_1'])   \n",
       "\n",
       "     elapsed_seconds  failed  index_size indexer_class  max_non_matches  \\\n",
       "483         0.069761   False      1686.0    BlockIndex              NaN   \n",
       "189         0.295529   False      5647.0    BlockIndex              NaN   \n",
       "204         0.337197   False     11812.0    BlockIndex              NaN   \n",
       "128         0.395938   False     34249.0    BlockIndex              NaN   \n",
       "187         0.320731   False      1979.0    BlockIndex              NaN   \n",
       "115         0.317954   False      1979.0    BlockIndex              NaN   \n",
       "127         0.317666   False      1979.0    BlockIndex              NaN   \n",
       "140         0.331411   False      1979.0    BlockIndex              NaN   \n",
       "220         0.309447   False      1979.0    BlockIndex              NaN   \n",
       "482         1.800290   False    389619.0    BlockIndex              NaN   \n",
       "239         0.370920   False     28400.0    BlockIndex              NaN   \n",
       "268         0.306980   False      1580.0    BlockIndex              NaN   \n",
       "196         0.508658   False     88351.0    BlockIndex              NaN   \n",
       "227         0.296237   False      1643.0    BlockIndex              NaN   \n",
       "212         1.211830   False    222207.0    BlockIndex              NaN   \n",
       "194         0.294836   False      1472.0    BlockIndex              NaN   \n",
       "489         0.102188   False     14052.0    BlockIndex              NaN   \n",
       "138         0.325827   False      1471.0    BlockIndex              NaN   \n",
       "210         0.312079   False      1466.0    BlockIndex              NaN   \n",
       "125         0.319502   False      1461.0    BlockIndex              NaN   \n",
       "225         0.298602   False      1456.0    BlockIndex              NaN   \n",
       "486         0.784496   False    149898.0    BlockIndex              NaN   \n",
       "260         0.297163   False      4060.0    BlockIndex              NaN   \n",
       "276         0.379718   False     34151.0    BlockIndex              NaN   \n",
       "258         0.079076   False      1287.0    BlockIndex              NaN   \n",
       "236         0.316014   False      1377.0    BlockIndex              NaN   \n",
       "192         0.312408   False      1658.0    BlockIndex              NaN   \n",
       "208         0.330807   False      2230.0    BlockIndex              NaN   \n",
       "223         0.325566   False      1265.0    BlockIndex              NaN   \n",
       "114         0.319703   False      1265.0    BlockIndex              NaN   \n",
       "270         0.324023   False      1258.0    BlockIndex              NaN   \n",
       "132         0.326778   False      5831.0    BlockIndex              NaN   \n",
       "241         0.325707   False      2686.0    BlockIndex              NaN   \n",
       "214         0.373399   False     12905.0    BlockIndex              NaN   \n",
       "488         0.315807   False     14837.0    BlockIndex              NaN   \n",
       "229         0.297485   False      1245.0    BlockIndex              NaN   \n",
       "102         0.350730   False      1245.0    BlockIndex              NaN   \n",
       "119         0.337206   False      2204.0    BlockIndex              NaN   \n",
       "135         0.327660   False      4117.0    BlockIndex              NaN   \n",
       "485         0.415172   False     48883.0    BlockIndex              NaN   \n",
       "250         0.315181   False     11780.0    BlockIndex              NaN   \n",
       "272         0.301735   False      1089.0    BlockIndex              NaN   \n",
       "262         0.302725   False      1245.0    BlockIndex              NaN   \n",
       "243         0.292492   False      1191.0    BlockIndex              NaN   \n",
       "216         0.301161   False      2108.0    BlockIndex              NaN   \n",
       "484         0.440356   False     49781.0    BlockIndex              NaN   \n",
       "199         0.323354   False      1461.0    BlockIndex              NaN   \n",
       "231         0.300556   False      1063.0    BlockIndex              NaN   \n",
       "248         0.299519   False      1036.0    BlockIndex              NaN   \n",
       "487         0.312787   False      6148.0    BlockIndex              NaN   \n",
       "\n",
       "     max_nulls  max_rank_differences  \\\n",
       "483        NaN                   NaN   \n",
       "189        NaN                   NaN   \n",
       "204        NaN                   NaN   \n",
       "128        NaN                   NaN   \n",
       "187        NaN                   NaN   \n",
       "115        NaN                   NaN   \n",
       "127        NaN                   NaN   \n",
       "140        NaN                   NaN   \n",
       "220        NaN                   NaN   \n",
       "482        NaN                   NaN   \n",
       "239        NaN                   NaN   \n",
       "268        NaN                   NaN   \n",
       "196        NaN                   NaN   \n",
       "227        NaN                   NaN   \n",
       "212        NaN                   NaN   \n",
       "194        NaN                   NaN   \n",
       "489        NaN                   NaN   \n",
       "138        NaN                   NaN   \n",
       "210        NaN                   NaN   \n",
       "125        NaN                   NaN   \n",
       "225        NaN                   NaN   \n",
       "486        NaN                   NaN   \n",
       "260        NaN                   NaN   \n",
       "276        NaN                   NaN   \n",
       "258        NaN                   NaN   \n",
       "236        NaN                   NaN   \n",
       "192        NaN                   NaN   \n",
       "208        NaN                   NaN   \n",
       "223        NaN                   NaN   \n",
       "114        NaN                   NaN   \n",
       "270        NaN                   NaN   \n",
       "132        NaN                   NaN   \n",
       "241        NaN                   NaN   \n",
       "214        NaN                   NaN   \n",
       "488        NaN                   NaN   \n",
       "229        NaN                   NaN   \n",
       "102        NaN                   NaN   \n",
       "119        NaN                   NaN   \n",
       "135        NaN                   NaN   \n",
       "485        NaN                   NaN   \n",
       "250        NaN                   NaN   \n",
       "272        NaN                   NaN   \n",
       "262        NaN                   NaN   \n",
       "243        NaN                   NaN   \n",
       "216        NaN                   NaN   \n",
       "484        NaN                   NaN   \n",
       "199        NaN                   NaN   \n",
       "231        NaN                   NaN   \n",
       "248        NaN                   NaN   \n",
       "487        NaN                   NaN   \n",
       "\n",
       "                                                    on  precision    recall  \\\n",
       "483                                       [soc_sec_id]   1.000000  0.871768   \n",
       "189            [date_of_birth_day, date_of_birth_year]   0.297149  0.867632   \n",
       "204          [date_of_birth_month, date_of_birth_year]   0.142059  0.867632   \n",
       "128           [date_of_birth_day, date_of_birth_month]   0.048907  0.866081   \n",
       "187            [date_of_birth_day, date_of_birth_orig]   0.843355  0.862978   \n",
       "115  [date_of_birth_day, date_of_birth_month, date_...   0.843355  0.862978   \n",
       "127  [date_of_birth_month, date_of_birth_orig, date...   0.843355  0.862978   \n",
       "140          [date_of_birth_month, date_of_birth_orig]   0.843355  0.862978   \n",
       "220           [date_of_birth_orig, date_of_birth_year]   0.843355  0.862978   \n",
       "482                                [date_of_birth_day]   0.004335  0.873320   \n",
       "239                        [date_of_birth_year, state]   0.056056  0.823164   \n",
       "268                                [soc_sec_id, state]   1.000000  0.816960   \n",
       "196                         [date_of_birth_day, state]   0.018019  0.823164   \n",
       "227                        [date_of_birth_orig, state]   0.957395  0.813340   \n",
       "212                       [date_of_birth_month, state]   0.007160  0.822647   \n",
       "194                    [date_of_birth_day, soc_sec_id]   1.000000  0.761117   \n",
       "489                                         [postcode]   0.104825  0.761634   \n",
       "138                   [date_of_birth_year, soc_sec_id]   1.000000  0.760600   \n",
       "210                  [date_of_birth_month, soc_sec_id]   1.000000  0.758014   \n",
       "125  [date_of_birth_month, date_of_birth_year, soc_...   1.000000  0.755429   \n",
       "225                   [date_of_birth_orig, soc_sec_id]   1.000000  0.752844   \n",
       "486                                    [street_number]   0.009687  0.750776   \n",
       "260                                  [postcode, state]   0.338916  0.711479   \n",
       "276                             [state, street_number]   0.039882  0.704240   \n",
       "258                             [postcode, soc_sec_id]   1.000000  0.665460   \n",
       "236                     [date_of_birth_year, postcode]   0.930283  0.662358   \n",
       "192                      [date_of_birth_day, postcode]   0.772014  0.661841   \n",
       "208                    [date_of_birth_month, postcode]   0.572646  0.660290   \n",
       "223                     [date_of_birth_orig, postcode]   1.000000  0.654085   \n",
       "114  [date_of_birth_month, date_of_birth_orig, date...   1.000000  0.654085   \n",
       "270                        [soc_sec_id, street_number]   1.000000  0.650465   \n",
       "132                 [date_of_birth_day, street_number]   0.215229  0.648914   \n",
       "241                [date_of_birth_year, street_number]   0.466121  0.647363   \n",
       "214               [date_of_birth_month, street_number]   0.097094  0.647880   \n",
       "488                                           [suburb]   0.084249  0.646329   \n",
       "229                [date_of_birth_orig, street_number]   0.992771  0.639090   \n",
       "102  [date_of_birth_day, date_of_birth_month, date_...   0.992771  0.639090   \n",
       "119          [date_of_birth_day, state, street_number]   0.537205  0.612203   \n",
       "135                                    [state, suburb]   0.285159  0.607032   \n",
       "485                                       [given_name]   0.024139  0.610134   \n",
       "250                                [given_name, state]   0.093973  0.572389   \n",
       "272                               [soc_sec_id, suburb]   1.000000  0.563082   \n",
       "262                          [postcode, street_number]   0.869076  0.559462   \n",
       "243                       [date_of_birth_year, suburb]   0.904282  0.556877   \n",
       "216                      [date_of_birth_month, suburb]   0.509962  0.555843   \n",
       "484                                          [surname]   0.021715  0.558945   \n",
       "199                        [date_of_birth_day, suburb]   0.733744  0.554292   \n",
       "231                       [date_of_birth_orig, suburb]   0.999059  0.549121   \n",
       "248                           [given_name, soc_sec_id]   1.000000  0.535677   \n",
       "487                                        [address_1]   0.167859  0.533609   \n",
       "\n",
       "     reduction_ratio  window renamed_indexer_class  recall+reduction  \n",
       "483         0.999865     NaN            BlockIndex          1.871633  \n",
       "189         0.999548     NaN            BlockIndex          1.867180  \n",
       "204         0.999055     NaN            BlockIndex          1.866687  \n",
       "128         0.997260     NaN            BlockIndex          1.863340  \n",
       "187         0.999842     NaN            BlockIndex          1.862820  \n",
       "115         0.999842     NaN            BlockIndex          1.862820  \n",
       "127         0.999842     NaN            BlockIndex          1.862820  \n",
       "140         0.999842     NaN            BlockIndex          1.862820  \n",
       "220         0.999842     NaN            BlockIndex          1.862820  \n",
       "482         0.968824     NaN            BlockIndex          1.842144  \n",
       "239         0.997728     NaN            BlockIndex          1.820892  \n",
       "268         0.999874     NaN            BlockIndex          1.816833  \n",
       "196         0.992931     NaN            BlockIndex          1.816095  \n",
       "227         0.999869     NaN            BlockIndex          1.813209  \n",
       "212         0.982220     NaN            BlockIndex          1.804867  \n",
       "194         0.999882     NaN            BlockIndex          1.760999  \n",
       "489         0.998876     NaN            BlockIndex          1.760510  \n",
       "138         0.999882     NaN            BlockIndex          1.760482  \n",
       "210         0.999883     NaN            BlockIndex          1.757897  \n",
       "125         0.999883     NaN            BlockIndex          1.755312  \n",
       "225         0.999883     NaN            BlockIndex          1.752727  \n",
       "486         0.988006     NaN            BlockIndex          1.738781  \n",
       "260         0.999675     NaN            BlockIndex          1.711154  \n",
       "276         0.997267     NaN            BlockIndex          1.701507  \n",
       "258         0.999897     NaN            BlockIndex          1.665357  \n",
       "236         0.999890     NaN            BlockIndex          1.662248  \n",
       "192         0.999867     NaN            BlockIndex          1.661708  \n",
       "208         0.999822     NaN            BlockIndex          1.660111  \n",
       "223         0.999899     NaN            BlockIndex          1.653984  \n",
       "114         0.999899     NaN            BlockIndex          1.653984  \n",
       "270         0.999899     NaN            BlockIndex          1.650365  \n",
       "132         0.999533     NaN            BlockIndex          1.648448  \n",
       "241         0.999785     NaN            BlockIndex          1.647148  \n",
       "214         0.998967     NaN            BlockIndex          1.646847  \n",
       "488         0.998813     NaN            BlockIndex          1.645142  \n",
       "229         0.999900     NaN            BlockIndex          1.638990  \n",
       "102         0.999900     NaN            BlockIndex          1.638990  \n",
       "119         0.999824     NaN            BlockIndex          1.612026  \n",
       "135         0.999671     NaN            BlockIndex          1.606703  \n",
       "485         0.996089     NaN            BlockIndex          1.606223  \n",
       "250         0.999057     NaN            BlockIndex          1.571446  \n",
       "272         0.999913     NaN            BlockIndex          1.562995  \n",
       "262         0.999900     NaN            BlockIndex          1.559363  \n",
       "243         0.999905     NaN            BlockIndex          1.556782  \n",
       "216         0.999831     NaN            BlockIndex          1.555674  \n",
       "484         0.996017     NaN            BlockIndex          1.554962  \n",
       "199         0.999883     NaN            BlockIndex          1.554175  \n",
       "231         0.999915     NaN            BlockIndex          1.549036  \n",
       "248         0.999917     NaN            BlockIndex          1.535594  \n",
       "487         0.999508     NaN            BlockIndex          1.533117  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats = get_indexing_stats('febrl2', cache_dir=cache_dir).copy()\n",
    "stats['renamed_indexer_class'] = stats['indexer_class']\n",
    "stats['renamed_indexer_class'][stats['renamed_indexer_class']=='AdjacentBlockIndex'] = 'NeighbourhoodBlockIndex'\n",
    "stats['recall+reduction'] = stats['recall'] + stats['reduction_ratio']\n",
    "#stats = stats[stats['max_non_matches'].isnull() | (stats['max_non_matches'] == 0)]\n",
    "#stats = stats[stats['on'].map(lambda x:(x is not np.nan) and {'soc_sec_id', 'date_of_birth_orig'} == set(x))]\n",
    "#stats = stats[stats['max_non_matches']>0]\n",
    "#stats = stats[(stats['recall']>0.5) & (stats['reduction_ratio']>0.5)]\n",
    "stats_ABI = stats[stats['renamed_indexer_class'] == 'BlockIndex']\n",
    "#stats_ABI = stats[stats['renamed_indexer_class'] == 'NeighbourhoodBlockIndex']\n",
    "#stats_ABI = stats[stats['renamed_indexer_class'] == 'SortedNeighbourhoodIndex']\n",
    "print(len(stats_ABI))\n",
    "stats_ABI.sort_values('recall+reduction', ascending=False).head(50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "cost_hull = ExtConvexHull(points = (1-stats_ABI[['reduction_ratio', 'recall']]).values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  0.00000000e+00,   1.00000000e+00],\n",
       "       [  3.11757552e-02,   1.26680455e-01],\n",
       "       [  1.34906981e-04,   1.28231644e-01]])"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cost_hull.points[cost_hull.minimizing_points()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbQAAAGfCAYAAAA3XyezAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nNABACAQNABDC/wOC8ZvvjcEtrgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f29919b2898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(7,7))\n",
    "ax=plt.gca()\n",
    "ax.set_xlim(0.5,1)\n",
    "ax.set_ylim(0.5,1)\n",
    "ax.text(0.75, 0.75, 'To be completed', ha='center', size=20)\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}