EVAL_GENDER.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "ecc39024",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import glob\n",
    "import os\n",
    "import numpy as np\n",
    "from pyannote.metrics.base import BaseMetric\n",
    "from inaGVAD.gender_metrics import WstpErr, IdentificationErrorRateLabel\n",
    "from pyannote.metrics.identification import IdentificationErrorRate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "989ecec9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pyannote.core import Annotation, Timeline, Segment\n",
    "\n",
    "def df2annot(df, col, rmnan=True, uri=None):\n",
    "    an = Annotation(uri=uri)\n",
    "#    print('df2annot', df)\n",
    "    for start, stop, val in zip(df.start, df.stop, df[col]):\n",
    "        if rmnan and val != val:\n",
    "            continue\n",
    "        seg = Segment(start, stop)\n",
    "        an[seg] = val\n",
    "    return an.support()\n",
    "\n",
    "\n",
    "def init_uem(df):\n",
    "    uem = Timeline()\n",
    "    uem.add(Segment(df.start[0], df.stop[len(df) - 1]))\n",
    "    return uem\n",
    "\n",
    "\n",
    "def rm_uem(uem, df, col, rmlist):\n",
    "    for start, stop, val in zip(df.start, df.stop, df[col]):\n",
    "        if val in rmlist:\n",
    "            uem = uem.extrude(Segment(start, stop))\n",
    "    return uem\n",
    "\n",
    "def keep_uem(uem, df, col, keeplist):\n",
    "    for start, stop, val in zip(df.start, df.stop, df[col]):\n",
    "        if val not in keeplist:\n",
    "            uem = uem.extrude(Segment(start, stop))\n",
    "    return uem\n",
    "\n",
    "\n",
    "class GenderEval:\n",
    "    def __init__(self, collar=.3):\n",
    "        self.wstp = WstpErr()\n",
    "        self.ier = IdentificationErrorRate(collar=collar)        \n",
    "        self.ierF = IdentificationErrorRateLabel('female', collar=collar)\n",
    "        self.ierM = IdentificationErrorRateLabel('male', collar=collar)\n",
    "        \n",
    "    def __call__(self, fref, fpred):\n",
    "        \n",
    "        uri, _ = os.path.splitext(os.path.basename(fref))\n",
    "        \n",
    "        # parse reference\n",
    "        dfref = pd.read_csv(fref)\n",
    "        anref = df2annot(dfref, 'speaker_gender', uri = uri)\n",
    "        uem = init_uem(dfref)\n",
    "        uem = rm_uem(uem, dfref, 'speaker_gender', ['undefgender'])\n",
    "        #print('main uem', uem)\n",
    "        #print(uem)\n",
    "        \n",
    "        # parse prediction\n",
    "        dfpred = pd.read_csv(fpred)\n",
    "        #print('dfpred', dfpred)\n",
    "        if len(dfpred) > 0:\n",
    "            dfpred = dfpred[dfpred.label.map(lambda x: x in ['male', 'female'])]\n",
    "        anpred = df2annot(dfpred, 'label', uri = uri)\n",
    "            \n",
    "        wstp = self.wstp(anref, anpred, uem=None)\n",
    "        \n",
    "        ier = self.ier(anref, anpred, uem=uem)\n",
    "        errF = self.ierF(anref,anpred,uem=uem)\n",
    "        errM = self.ierM(anref,anpred,uem=uem)\n",
    "        \n",
    "        return wstp, ier, errF, errM\n",
    "    \n",
    "    def reset(self):\n",
    "        self.wstp.reset()\n",
    "        self.ier.reset()\n",
    "        self.ierF.reset()\n",
    "        self.ierM.reset()\n",
    "        \n",
    "    def report(self):\n",
    "        wstp = self.wstp.report()\n",
    "        wstp.columns = [('WSTP', e1, e2) for e1, e2 in wstp.columns]\n",
    "        ier = self.ier.report()\n",
    "        ier.columns = [('IER', e1, e2) for e1, e2 in ier.columns]\n",
    "        ierF = self.ierF.report()\n",
    "        ierF.columns = [('IER Female', e1, e2) for e1, e2 in ierF.columns]\n",
    "        ierM = self.ierM.report()\n",
    "        ierM.columns = [('IER Male', e1, e2) for e1, e2 in ierM.columns]\n",
    "        return wstp.join(ier).join(ierF).join(ierM)\n",
    "        \n",
    "#        if onlypercent:\n",
    "#            ret = ret[[k for k in ret if k[2] == '%']]\n",
    "#            ret.columns = [(e1, e2) for (e1, e2, e3) in ret]\n",
    "#        return ret\n",
    "\n",
    "    def compare_lfiles(self, lref, lhyp, simplify=None):\n",
    "#        assert(simplify in [None, 'onlypercent', 'genderdetail', 'gender', ''])\n",
    "\n",
    "        for ref, hyp in zip(lref, lhyp):\n",
    "            self(ref, hyp)\n",
    "        ret = self.report()\n",
    "        self.reset()\n",
    "        return ret\n",
    "        \n",
    "#        if simplify == 'only percent':\n",
    "#            ret = ret[[k for k in ret if k[2] == '%']]\n",
    "#            ret.columns = [(e1, e2) for (e1, e2, e3) in ret]\n",
    "\n",
    "    \n",
    "    def compare_category(self, pred_dir, criterion, csvfname='./annotations/filesplit/testset.csv'):\n",
    "\n",
    "        df = pd.read_csv(csvfname)\n",
    "        lret = []\n",
    "        #ldf = []\n",
    "        #dret = {}\n",
    "        ltotal = []\n",
    "        \n",
    "        for k, sdf in df.groupby(criterion):\n",
    "            src = sdf.fileid.map(lambda x: './annotations/detailed_csv/%s.csv' % x)\n",
    "            dst = sdf.fileid.map(lambda x: '%s/%s.csv' % (pred_dir, x))\n",
    "            \n",
    "            ret = self.compare_lfiles(src, dst)\n",
    "\n",
    "            # remove total from result\n",
    "            dtotal = ret.tail(1).to_dict(orient='records')[0]    \n",
    "            ret = ret[:(len(ret) - 1)]\n",
    "            lret.append(ret)\n",
    "  \n",
    "            if isinstance(criterion, list):\n",
    "                for i, key in enumerate(criterion):\n",
    "                    dtotal[key] = k[i]\n",
    "            else:\n",
    "                dtotal[criterion] = k\n",
    "\n",
    "            dtotal['nbfile'] = len(ret)\n",
    "            wstp = ret[('WSTP', 'WSTP error', '%')]\n",
    "            dtotal['WSTP_RMSE'] = np.sqrt(np.mean(wstp ** 2))\n",
    "            \n",
    "            ltotal.append(dtotal)\n",
    "\n",
    "        details = df.join(pd.concat(lret), on='fileid')\n",
    "        return details, pd.DataFrame.from_dict(ltotal).set_index(criterion)\n",
    "\n",
    "def tuple3cols(df):\n",
    "    return [k for k in df.columns if isinstance(k, tuple) and len(k) == 3]\n",
    "\n",
    "def simplify_results(df, add_speech_percent=True, rmdur=True, iergbias = True, ierdetail=False, genderdetail=False):\n",
    "    #assert strategy in ['percents', 'wstp_ier', 'wstp_iergender', 'wstp_iergenderdetail']\n",
    "    \n",
    "#    if add_speech_percent:\n",
    "#        \n",
    "#        if 'nbfile' in df:\n",
    "#            nbfile = df.nbfile\n",
    "#            df = df.drop(['nbfile'], axis=1)\n",
    "        \n",
    "    ldrop = []\n",
    "    if rmdur:\n",
    "        ldrop += [k for k in df.columns if isinstance(k, tuple) and len(k) == 3 and k[2] != '%']\n",
    "        \n",
    "    if iergbias:\n",
    "        df['IER gender bias'] = df[('IER Female', 'identification error rate', '%')] - df[('IER Male', 'identification error rate', '%')]\n",
    "        \n",
    "    if ierdetail == False:\n",
    "        ldrop += [k for k in tuple3cols(df) if k[0] == 'IER' and k[1] != 'identification error rate']\n",
    "\n",
    "    if genderdetail == False:\n",
    "        ldrop += [k for k in tuple3cols(df) if k[0] in ['IER Female', 'IER Male'] and k[1] != 'identification error rate']\n",
    "        \n",
    "    return df.drop(ldrop, axis=1)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "7b0783be",
   "metadata": {},
   "outputs": [],
   "source": [
    "ref = './annotations/detailed_csv/'\n",
    "lref = sorted(glob.glob('./annotations/detailed_csv/*.csv'))\n",
    "iss = './automatic_baselines/inaspeechsegmenter/'\n",
    "liss = sorted(glob.glob('./automatic_baselines/inaspeechsegmenter/*.csv'))\n",
    "lium = './automatic_baselines/liumspkdiarization_csv/'\n",
    "llium = sorted(glob.glob('./automatic_baselines/liumspkdiarization_csv/*.csv'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "53a17ead",
   "metadata": {},
   "outputs": [],
   "source": [
    "# init evaluator and COLLAR (zone to ignore around boundaries)\n",
    "ge = GenderEval(collar=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "45941f56",
   "metadata": {},
   "outputs": [],
   "source": [
    "# evaluation performed on dev split : return results per file in details and per category in category\n",
    "details, categories = ge.compare_category(iss, ['channel_category'])#, 'vad_difficulty'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "694f20b3",
   "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>(WSTP, WSTP error, %)</th>\n",
       "      <th>(IER, identification error rate, %)</th>\n",
       "      <th>(IER Female, identification error rate, %)</th>\n",
       "      <th>(IER Male, identification error rate, %)</th>\n",
       "      <th>nbfile</th>\n",
       "      <th>WSTP_RMSE</th>\n",
       "      <th>IER gender bias</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>channel_category</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>generalist_radio</th>\n",
       "      <td>0.670746</td>\n",
       "      <td>5.851559</td>\n",
       "      <td>6.088145</td>\n",
       "      <td>5.800402</td>\n",
       "      <td>34</td>\n",
       "      <td>4.434916</td>\n",
       "      <td>0.287743</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>generalist_tv</th>\n",
       "      <td>-1.437044</td>\n",
       "      <td>25.517894</td>\n",
       "      <td>26.107704</td>\n",
       "      <td>25.120865</td>\n",
       "      <td>93</td>\n",
       "      <td>16.602669</td>\n",
       "      <td>0.986839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>music_radio</th>\n",
       "      <td>3.610936</td>\n",
       "      <td>16.286951</td>\n",
       "      <td>5.260995</td>\n",
       "      <td>23.403875</td>\n",
       "      <td>46</td>\n",
       "      <td>15.770984</td>\n",
       "      <td>-18.142881</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>news_tv</th>\n",
       "      <td>1.042652</td>\n",
       "      <td>6.466632</td>\n",
       "      <td>5.425766</td>\n",
       "      <td>7.250946</td>\n",
       "      <td>44</td>\n",
       "      <td>4.254993</td>\n",
       "      <td>-1.825180</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  (WSTP, WSTP error, %)  (IER, identification error rate, %)  \\\n",
       "channel_category                                                               \n",
       "generalist_radio               0.670746                             5.851559   \n",
       "generalist_tv                 -1.437044                            25.517894   \n",
       "music_radio                    3.610936                            16.286951   \n",
       "news_tv                        1.042652                             6.466632   \n",
       "\n",
       "                  (IER Female, identification error rate, %)  \\\n",
       "channel_category                                               \n",
       "generalist_radio                                    6.088145   \n",
       "generalist_tv                                      26.107704   \n",
       "music_radio                                         5.260995   \n",
       "news_tv                                             5.425766   \n",
       "\n",
       "                  (IER Male, identification error rate, %)  nbfile  WSTP_RMSE  \\\n",
       "channel_category                                                                \n",
       "generalist_radio                                  5.800402      34   4.434916   \n",
       "generalist_tv                                    25.120865      93  16.602669   \n",
       "music_radio                                      23.403875      46  15.770984   \n",
       "news_tv                                           7.250946      44   4.254993   \n",
       "\n",
       "                  IER gender bias  \n",
       "channel_category                   \n",
       "generalist_radio         0.287743  \n",
       "generalist_tv            0.986839  \n",
       "music_radio            -18.142881  \n",
       "news_tv                 -1.825180  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# human readable simplication of results per category\n",
    "simplify_results(categories)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "884466d7",
   "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>fileid</th>\n",
       "      <th>media</th>\n",
       "      <th>channel_code</th>\n",
       "      <th>channel_name</th>\n",
       "      <th>channel_category</th>\n",
       "      <th>broadcast_hour</th>\n",
       "      <th>vad_difficulty</th>\n",
       "      <th>(WSTP, WSTP error, %)</th>\n",
       "      <th>(IER, identification error rate, %)</th>\n",
       "      <th>(IER Female, identification error rate, %)</th>\n",
       "      <th>(IER Male, identification error rate, %)</th>\n",
       "      <th>IER gender bias</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>radio-FBL-011904</td>\n",
       "      <td>radio</td>\n",
       "      <td>FBL</td>\n",
       "      <td>France Bleu</td>\n",
       "      <td>music_radio</td>\n",
       "      <td>1</td>\n",
       "      <td>EASY</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>radio-FBL-025455</td>\n",
       "      <td>radio</td>\n",
       "      <td>FBL</td>\n",
       "      <td>France Bleu</td>\n",
       "      <td>music_radio</td>\n",
       "      <td>2</td>\n",
       "      <td>MEDIUM</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>radio-FBL-035452</td>\n",
       "      <td>radio</td>\n",
       "      <td>FBL</td>\n",
       "      <td>France Bleu</td>\n",
       "      <td>music_radio</td>\n",
       "      <td>3</td>\n",
       "      <td>EASY</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>radio-FBL-044522</td>\n",
       "      <td>radio</td>\n",
       "      <td>FBL</td>\n",
       "      <td>France Bleu</td>\n",
       "      <td>music_radio</td>\n",
       "      <td>4</td>\n",
       "      <td>MEDIUM</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>radio-FBL-051706</td>\n",
       "      <td>radio</td>\n",
       "      <td>FBL</td>\n",
       "      <td>France Bleu</td>\n",
       "      <td>music_radio</td>\n",
       "      <td>5</td>\n",
       "      <td>EASY</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.633636</td>\n",
       "      <td>2.633636</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.633636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>212</th>\n",
       "      <td>tv-TF1-194413</td>\n",
       "      <td>tv</td>\n",
       "      <td>TF1</td>\n",
       "      <td>TF1</td>\n",
       "      <td>generalist_tv</td>\n",
       "      <td>19</td>\n",
       "      <td>HARD</td>\n",
       "      <td>7.479113</td>\n",
       "      <td>77.934416</td>\n",
       "      <td>41.005938</td>\n",
       "      <td>100.189447</td>\n",
       "      <td>-59.183509</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>213</th>\n",
       "      <td>tv-TF1-215118</td>\n",
       "      <td>tv</td>\n",
       "      <td>TF1</td>\n",
       "      <td>TF1</td>\n",
       "      <td>generalist_tv</td>\n",
       "      <td>21</td>\n",
       "      <td>HARD</td>\n",
       "      <td>-4.688474</td>\n",
       "      <td>10.443075</td>\n",
       "      <td>14.007484</td>\n",
       "      <td>2.263856</td>\n",
       "      <td>11.743628</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>214</th>\n",
       "      <td>tv-TF1-225026</td>\n",
       "      <td>tv</td>\n",
       "      <td>TF1</td>\n",
       "      <td>TF1</td>\n",
       "      <td>generalist_tv</td>\n",
       "      <td>22</td>\n",
       "      <td>HARD</td>\n",
       "      <td>-5.115000</td>\n",
       "      <td>32.358499</td>\n",
       "      <td>63.022585</td>\n",
       "      <td>25.079845</td>\n",
       "      <td>37.942740</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>215</th>\n",
       "      <td>tv-TFX-230837</td>\n",
       "      <td>tv</td>\n",
       "      <td>TFX</td>\n",
       "      <td>TFX</td>\n",
       "      <td>generalist_tv</td>\n",
       "      <td>23</td>\n",
       "      <td>HARD</td>\n",
       "      <td>25.488273</td>\n",
       "      <td>81.186036</td>\n",
       "      <td>46.626476</td>\n",
       "      <td>1745.129225</td>\n",
       "      <td>-1698.502748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>216</th>\n",
       "      <td>tv-TFX-231251</td>\n",
       "      <td>tv</td>\n",
       "      <td>TFX</td>\n",
       "      <td>TFX</td>\n",
       "      <td>generalist_tv</td>\n",
       "      <td>23</td>\n",
       "      <td>MEDIUM</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>95.639871</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>95.639871</td>\n",
       "      <td>-95.639871</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>217 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "               fileid  media channel_code channel_name channel_category  \\\n",
       "0    radio-FBL-011904  radio          FBL  France Bleu      music_radio   \n",
       "1    radio-FBL-025455  radio          FBL  France Bleu      music_radio   \n",
       "2    radio-FBL-035452  radio          FBL  France Bleu      music_radio   \n",
       "3    radio-FBL-044522  radio          FBL  France Bleu      music_radio   \n",
       "4    radio-FBL-051706  radio          FBL  France Bleu      music_radio   \n",
       "..                ...    ...          ...          ...              ...   \n",
       "212     tv-TF1-194413     tv          TF1          TF1    generalist_tv   \n",
       "213     tv-TF1-215118     tv          TF1          TF1    generalist_tv   \n",
       "214     tv-TF1-225026     tv          TF1          TF1    generalist_tv   \n",
       "215     tv-TFX-230837     tv          TFX          TFX    generalist_tv   \n",
       "216     tv-TFX-231251     tv          TFX          TFX    generalist_tv   \n",
       "\n",
       "     broadcast_hour vad_difficulty  (WSTP, WSTP error, %)  \\\n",
       "0                 1           EASY               0.000000   \n",
       "1                 2         MEDIUM               0.000000   \n",
       "2                 3           EASY               0.000000   \n",
       "3                 4         MEDIUM               0.000000   \n",
       "4                 5           EASY               0.000000   \n",
       "..              ...            ...                    ...   \n",
       "212              19           HARD               7.479113   \n",
       "213              21           HARD              -4.688474   \n",
       "214              22           HARD              -5.115000   \n",
       "215              23           HARD              25.488273   \n",
       "216              23         MEDIUM               0.000000   \n",
       "\n",
       "     (IER, identification error rate, %)  \\\n",
       "0                               0.000000   \n",
       "1                               0.000000   \n",
       "2                               0.000000   \n",
       "3                               0.000000   \n",
       "4                               2.633636   \n",
       "..                                   ...   \n",
       "212                            77.934416   \n",
       "213                            10.443075   \n",
       "214                            32.358499   \n",
       "215                            81.186036   \n",
       "216                            95.639871   \n",
       "\n",
       "     (IER Female, identification error rate, %)  \\\n",
       "0                                      0.000000   \n",
       "1                                      0.000000   \n",
       "2                                      0.000000   \n",
       "3                                      0.000000   \n",
       "4                                      2.633636   \n",
       "..                                          ...   \n",
       "212                                   41.005938   \n",
       "213                                   14.007484   \n",
       "214                                   63.022585   \n",
       "215                                   46.626476   \n",
       "216                                    0.000000   \n",
       "\n",
       "     (IER Male, identification error rate, %)  IER gender bias  \n",
       "0                                    0.000000         0.000000  \n",
       "1                                    0.000000         0.000000  \n",
       "2                                    0.000000         0.000000  \n",
       "3                                    0.000000         0.000000  \n",
       "4                                    0.000000         2.633636  \n",
       "..                                        ...              ...  \n",
       "212                                100.189447       -59.183509  \n",
       "213                                  2.263856        11.743628  \n",
       "214                                 25.079845        37.942740  \n",
       "215                               1745.129225     -1698.502748  \n",
       "216                                 95.639871       -95.639871  \n",
       "\n",
       "[217 rows x 12 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# simplified view of results per file\n",
    "simplify_results(details)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "f9bba7dd",
   "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>(WSTP, WSTP error, %)</th>\n",
       "      <th>(WSTP, ref_male_dur, )</th>\n",
       "      <th>(WSTP, ref_female_dur, )</th>\n",
       "      <th>(WSTP, hyp_male_dur, )</th>\n",
       "      <th>(WSTP, hyp_female_dur, )</th>\n",
       "      <th>(IER, identification error rate, %)</th>\n",
       "      <th>(IER, total, )</th>\n",
       "      <th>(IER, correct, )</th>\n",
       "      <th>(IER, correct, %)</th>\n",
       "      <th>(IER, false alarm, )</th>\n",
       "      <th>...</th>\n",
       "      <th>(IER Male, correct, %)</th>\n",
       "      <th>(IER Male, false alarm, )</th>\n",
       "      <th>(IER Male, false alarm, %)</th>\n",
       "      <th>(IER Male, missed detection, )</th>\n",
       "      <th>(IER Male, missed detection, %)</th>\n",
       "      <th>(IER Male, confusion, )</th>\n",
       "      <th>(IER Male, confusion, %)</th>\n",
       "      <th>nbfile</th>\n",
       "      <th>WSTP_RMSE</th>\n",
       "      <th>IER gender bias</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>channel_category</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>generalist_radio</th>\n",
       "      <td>0.670746</td>\n",
       "      <td>1491.8560</td>\n",
       "      <td>334.2150</td>\n",
       "      <td>1630.30</td>\n",
       "      <td>348.98</td>\n",
       "      <td>5.851559</td>\n",
       "      <td>1699.547</td>\n",
       "      <td>1681.382</td>\n",
       "      <td>98.931186</td>\n",
       "      <td>81.285</td>\n",
       "      <td>...</td>\n",
       "      <td>99.371827</td>\n",
       "      <td>67.717</td>\n",
       "      <td>4.845977</td>\n",
       "      <td>3.950</td>\n",
       "      <td>0.282671</td>\n",
       "      <td>9.387</td>\n",
       "      <td>0.671754</td>\n",
       "      <td>34</td>\n",
       "      <td>4.434916</td>\n",
       "      <td>0.287743</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>generalist_tv</th>\n",
       "      <td>-1.437044</td>\n",
       "      <td>1971.8200</td>\n",
       "      <td>1367.6720</td>\n",
       "      <td>2158.52</td>\n",
       "      <td>1588.36</td>\n",
       "      <td>25.517894</td>\n",
       "      <td>2823.646</td>\n",
       "      <td>2496.775</td>\n",
       "      <td>88.423797</td>\n",
       "      <td>393.664</td>\n",
       "      <td>...</td>\n",
       "      <td>87.222932</td>\n",
       "      <td>272.873</td>\n",
       "      <td>16.169054</td>\n",
       "      <td>103.127</td>\n",
       "      <td>6.110777</td>\n",
       "      <td>47.946</td>\n",
       "      <td>2.841034</td>\n",
       "      <td>93</td>\n",
       "      <td>16.602669</td>\n",
       "      <td>0.986839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>music_radio</th>\n",
       "      <td>3.610936</td>\n",
       "      <td>218.5525</td>\n",
       "      <td>141.8625</td>\n",
       "      <td>265.16</td>\n",
       "      <td>147.54</td>\n",
       "      <td>16.286951</td>\n",
       "      <td>322.522</td>\n",
       "      <td>313.365</td>\n",
       "      <td>97.160814</td>\n",
       "      <td>43.372</td>\n",
       "      <td>...</td>\n",
       "      <td>98.655653</td>\n",
       "      <td>40.980</td>\n",
       "      <td>20.907523</td>\n",
       "      <td>2.635</td>\n",
       "      <td>1.344347</td>\n",
       "      <td>2.258</td>\n",
       "      <td>1.152006</td>\n",
       "      <td>46</td>\n",
       "      <td>15.770984</td>\n",
       "      <td>-18.142881</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>news_tv</th>\n",
       "      <td>1.042652</td>\n",
       "      <td>1287.6910</td>\n",
       "      <td>975.5460</td>\n",
       "      <td>1437.84</td>\n",
       "      <td>1043.82</td>\n",
       "      <td>6.466632</td>\n",
       "      <td>2042.114</td>\n",
       "      <td>2019.279</td>\n",
       "      <td>98.881796</td>\n",
       "      <td>109.221</td>\n",
       "      <td>...</td>\n",
       "      <td>99.406309</td>\n",
       "      <td>66.289</td>\n",
       "      <td>5.692100</td>\n",
       "      <td>6.871</td>\n",
       "      <td>0.589999</td>\n",
       "      <td>11.283</td>\n",
       "      <td>0.968848</td>\n",
       "      <td>44</td>\n",
       "      <td>4.254993</td>\n",
       "      <td>-1.825180</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>4 rows × 38 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                  (WSTP, WSTP error, %)  (WSTP, ref_male_dur, )  \\\n",
       "channel_category                                                  \n",
       "generalist_radio               0.670746               1491.8560   \n",
       "generalist_tv                 -1.437044               1971.8200   \n",
       "music_radio                    3.610936                218.5525   \n",
       "news_tv                        1.042652               1287.6910   \n",
       "\n",
       "                  (WSTP, ref_female_dur, )  (WSTP, hyp_male_dur, )  \\\n",
       "channel_category                                                     \n",
       "generalist_radio                  334.2150                 1630.30   \n",
       "generalist_tv                    1367.6720                 2158.52   \n",
       "music_radio                       141.8625                  265.16   \n",
       "news_tv                           975.5460                 1437.84   \n",
       "\n",
       "                  (WSTP, hyp_female_dur, )  \\\n",
       "channel_category                             \n",
       "generalist_radio                    348.98   \n",
       "generalist_tv                      1588.36   \n",
       "music_radio                         147.54   \n",
       "news_tv                            1043.82   \n",
       "\n",
       "                  (IER, identification error rate, %)  (IER, total, )  \\\n",
       "channel_category                                                        \n",
       "generalist_radio                             5.851559        1699.547   \n",
       "generalist_tv                               25.517894        2823.646   \n",
       "music_radio                                 16.286951         322.522   \n",
       "news_tv                                      6.466632        2042.114   \n",
       "\n",
       "                  (IER, correct, )  (IER, correct, %)  (IER, false alarm, )  \\\n",
       "channel_category                                                              \n",
       "generalist_radio          1681.382          98.931186                81.285   \n",
       "generalist_tv             2496.775          88.423797               393.664   \n",
       "music_radio                313.365          97.160814                43.372   \n",
       "news_tv                   2019.279          98.881796               109.221   \n",
       "\n",
       "                  ...  (IER Male, correct, %)  (IER Male, false alarm, )  \\\n",
       "channel_category  ...                                                      \n",
       "generalist_radio  ...               99.371827                     67.717   \n",
       "generalist_tv     ...               87.222932                    272.873   \n",
       "music_radio       ...               98.655653                     40.980   \n",
       "news_tv           ...               99.406309                     66.289   \n",
       "\n",
       "                  (IER Male, false alarm, %)  (IER Male, missed detection, )  \\\n",
       "channel_category                                                               \n",
       "generalist_radio                    4.845977                           3.950   \n",
       "generalist_tv                      16.169054                         103.127   \n",
       "music_radio                        20.907523                           2.635   \n",
       "news_tv                             5.692100                           6.871   \n",
       "\n",
       "                  (IER Male, missed detection, %)  (IER Male, confusion, )  \\\n",
       "channel_category                                                             \n",
       "generalist_radio                         0.282671                    9.387   \n",
       "generalist_tv                            6.110777                   47.946   \n",
       "music_radio                              1.344347                    2.258   \n",
       "news_tv                                  0.589999                   11.283   \n",
       "\n",
       "                  (IER Male, confusion, %)  nbfile  WSTP_RMSE  IER gender bias  \n",
       "channel_category                                                                \n",
       "generalist_radio                  0.671754      34   4.434916         0.287743  \n",
       "generalist_tv                     2.841034      93  16.602669         0.986839  \n",
       "music_radio                       1.152006      46  15.770984       -18.142881  \n",
       "news_tv                           0.968848      44   4.254993        -1.825180  \n",
       "\n",
       "[4 rows x 38 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# detailled metrics on categories (hard 2 read)\n",
    "categories\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "ffaf4f83",
   "metadata": {},
   "outputs": [],
   "source": [
    "# LIUM results are les convincing\n",
    "details, categories = ge.compare_category(lium, ['channel_category'])#, 'vad_difficulty'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "dc70d43e",
   "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>(WSTP, WSTP error, %)</th>\n",
       "      <th>(IER, identification error rate, %)</th>\n",
       "      <th>(IER Female, identification error rate, %)</th>\n",
       "      <th>(IER Male, identification error rate, %)</th>\n",
       "      <th>nbfile</th>\n",
       "      <th>WSTP_RMSE</th>\n",
       "      <th>IER gender bias</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>channel_category</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>generalist_radio</th>\n",
       "      <td>-2.404443</td>\n",
       "      <td>9.757482</td>\n",
       "      <td>27.929481</td>\n",
       "      <td>5.828096</td>\n",
       "      <td>34</td>\n",
       "      <td>11.288720</td>\n",
       "      <td>22.101385</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>generalist_tv</th>\n",
       "      <td>-10.125963</td>\n",
       "      <td>47.260492</td>\n",
       "      <td>70.070008</td>\n",
       "      <td>31.906318</td>\n",
       "      <td>93</td>\n",
       "      <td>32.591160</td>\n",
       "      <td>38.163689</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>music_radio</th>\n",
       "      <td>-3.552042</td>\n",
       "      <td>219.117456</td>\n",
       "      <td>247.118151</td>\n",
       "      <td>201.043846</td>\n",
       "      <td>46</td>\n",
       "      <td>36.797889</td>\n",
       "      <td>46.074305</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>news_tv</th>\n",
       "      <td>-2.053974</td>\n",
       "      <td>15.927465</td>\n",
       "      <td>21.493958</td>\n",
       "      <td>11.732995</td>\n",
       "      <td>44</td>\n",
       "      <td>20.654653</td>\n",
       "      <td>9.760962</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  (WSTP, WSTP error, %)  (IER, identification error rate, %)  \\\n",
       "channel_category                                                               \n",
       "generalist_radio              -2.404443                             9.757482   \n",
       "generalist_tv                -10.125963                            47.260492   \n",
       "music_radio                   -3.552042                           219.117456   \n",
       "news_tv                       -2.053974                            15.927465   \n",
       "\n",
       "                  (IER Female, identification error rate, %)  \\\n",
       "channel_category                                               \n",
       "generalist_radio                                   27.929481   \n",
       "generalist_tv                                      70.070008   \n",
       "music_radio                                       247.118151   \n",
       "news_tv                                            21.493958   \n",
       "\n",
       "                  (IER Male, identification error rate, %)  nbfile  WSTP_RMSE  \\\n",
       "channel_category                                                                \n",
       "generalist_radio                                  5.828096      34  11.288720   \n",
       "generalist_tv                                    31.906318      93  32.591160   \n",
       "music_radio                                     201.043846      46  36.797889   \n",
       "news_tv                                          11.732995      44  20.654653   \n",
       "\n",
       "                  IER gender bias  \n",
       "channel_category                   \n",
       "generalist_radio        22.101385  \n",
       "generalist_tv           38.163689  \n",
       "music_radio             46.074305  \n",
       "news_tv                  9.760962  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "simplify_results(categories)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "80ca64ce",
   "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>(WSTP, WSTP error, %)</th>\n",
       "      <th>(IER, identification error rate, %)</th>\n",
       "      <th>(IER Female, identification error rate, %)</th>\n",
       "      <th>(IER Male, identification error rate, %)</th>\n",
       "      <th>nbfile</th>\n",
       "      <th>WSTP_RMSE</th>\n",
       "      <th>IER gender bias</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>vad_difficulty</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>EASY</th>\n",
       "      <td>-0.829049</td>\n",
       "      <td>4.933390</td>\n",
       "      <td>7.175926</td>\n",
       "      <td>3.582428</td>\n",
       "      <td>95</td>\n",
       "      <td>8.963020</td>\n",
       "      <td>3.593498</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HARD</th>\n",
       "      <td>-0.819120</td>\n",
       "      <td>55.856100</td>\n",
       "      <td>46.418721</td>\n",
       "      <td>62.566244</td>\n",
       "      <td>44</td>\n",
       "      <td>22.436811</td>\n",
       "      <td>-16.147523</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MEDIUM</th>\n",
       "      <td>0.929599</td>\n",
       "      <td>15.962817</td>\n",
       "      <td>16.286284</td>\n",
       "      <td>15.818216</td>\n",
       "      <td>78</td>\n",
       "      <td>10.597541</td>\n",
       "      <td>0.468069</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                (WSTP, WSTP error, %)  (IER, identification error rate, %)  \\\n",
       "vad_difficulty                                                               \n",
       "EASY                        -0.829049                             4.933390   \n",
       "HARD                        -0.819120                            55.856100   \n",
       "MEDIUM                       0.929599                            15.962817   \n",
       "\n",
       "                (IER Female, identification error rate, %)  \\\n",
       "vad_difficulty                                               \n",
       "EASY                                              7.175926   \n",
       "HARD                                             46.418721   \n",
       "MEDIUM                                           16.286284   \n",
       "\n",
       "                (IER Male, identification error rate, %)  nbfile  WSTP_RMSE  \\\n",
       "vad_difficulty                                                                \n",
       "EASY                                            3.582428      95   8.963020   \n",
       "HARD                                           62.566244      44  22.436811   \n",
       "MEDIUM                                         15.818216      78  10.597541   \n",
       "\n",
       "                IER gender bias  \n",
       "vad_difficulty                   \n",
       "EASY                   3.593498  \n",
       "HARD                 -16.147523  \n",
       "MEDIUM                 0.468069  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#ina speech segmenter again : harder VAD means strong gender biases and low IER\n",
    "details, categories = ge.compare_category(iss, ['vad_difficulty'])\n",
    "simplify_results(categories)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "ba1ba3ac",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>(WSTP, WSTP error, %)</th>\n",
       "      <th>(IER, identification error rate, %)</th>\n",
       "      <th>(IER Female, identification error rate, %)</th>\n",
       "      <th>(IER Male, identification error rate, %)</th>\n",
       "      <th>nbfile</th>\n",
       "      <th>WSTP_RMSE</th>\n",
       "      <th>IER gender bias</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>channel_category</th>\n",
       "      <th>vad_difficulty</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">generalist_radio</th>\n",
       "      <th>EASY</th>\n",
       "      <td>1.016144</td>\n",
       "      <td>2.432442</td>\n",
       "      <td>1.714421</td>\n",
       "      <td>2.611223</td>\n",
       "      <td>25</td>\n",
       "      <td>3.025124</td>\n",
       "      <td>-0.896802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HARD</th>\n",
       "      <td>2.540544</td>\n",
       "      <td>35.483808</td>\n",
       "      <td>30.055415</td>\n",
       "      <td>37.800322</td>\n",
       "      <td>3</td>\n",
       "      <td>9.663539</td>\n",
       "      <td>-7.744907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MEDIUM</th>\n",
       "      <td>-1.451433</td>\n",
       "      <td>11.638845</td>\n",
       "      <td>77.481359</td>\n",
       "      <td>10.105248</td>\n",
       "      <td>6</td>\n",
       "      <td>5.160619</td>\n",
       "      <td>67.376111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">generalist_tv</th>\n",
       "      <th>EASY</th>\n",
       "      <td>-6.609124</td>\n",
       "      <td>15.989948</td>\n",
       "      <td>23.455468</td>\n",
       "      <td>8.103532</td>\n",
       "      <td>17</td>\n",
       "      <td>20.367249</td>\n",
       "      <td>15.351936</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HARD</th>\n",
       "      <td>-3.376755</td>\n",
       "      <td>54.486051</td>\n",
       "      <td>48.741820</td>\n",
       "      <td>58.900198</td>\n",
       "      <td>27</td>\n",
       "      <td>19.679952</td>\n",
       "      <td>-10.158377</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MEDIUM</th>\n",
       "      <td>0.869468</td>\n",
       "      <td>18.293169</td>\n",
       "      <td>17.229385</td>\n",
       "      <td>18.868623</td>\n",
       "      <td>49</td>\n",
       "      <td>12.877909</td>\n",
       "      <td>-1.639238</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">music_radio</th>\n",
       "      <th>EASY</th>\n",
       "      <td>-2.263857</td>\n",
       "      <td>2.545830</td>\n",
       "      <td>2.347034</td>\n",
       "      <td>2.679484</td>\n",
       "      <td>25</td>\n",
       "      <td>2.069831</td>\n",
       "      <td>-0.332450</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HARD</th>\n",
       "      <td>71.434134</td>\n",
       "      <td>652.633362</td>\n",
       "      <td>62.151487</td>\n",
       "      <td>2240.037477</td>\n",
       "      <td>13</td>\n",
       "      <td>29.516576</td>\n",
       "      <td>-2177.885990</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MEDIUM</th>\n",
       "      <td>4.543460</td>\n",
       "      <td>10.021808</td>\n",
       "      <td>6.496364</td>\n",
       "      <td>11.822060</td>\n",
       "      <td>8</td>\n",
       "      <td>1.014231</td>\n",
       "      <td>-5.325696</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">news_tv</th>\n",
       "      <th>EASY</th>\n",
       "      <td>0.435185</td>\n",
       "      <td>3.082951</td>\n",
       "      <td>2.725143</td>\n",
       "      <td>3.419553</td>\n",
       "      <td>28</td>\n",
       "      <td>2.951893</td>\n",
       "      <td>-0.694410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HARD</th>\n",
       "      <td>-9.318224</td>\n",
       "      <td>32.701088</td>\n",
       "      <td>28.590344</td>\n",
       "      <td>35.058109</td>\n",
       "      <td>1</td>\n",
       "      <td>9.318224</td>\n",
       "      <td>-6.467765</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MEDIUM</th>\n",
       "      <td>1.590077</td>\n",
       "      <td>12.546062</td>\n",
       "      <td>13.098857</td>\n",
       "      <td>12.290901</td>\n",
       "      <td>15</td>\n",
       "      <td>5.572590</td>\n",
       "      <td>0.807956</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 (WSTP, WSTP error, %)  \\\n",
       "channel_category vad_difficulty                          \n",
       "generalist_radio EASY                         1.016144   \n",
       "                 HARD                         2.540544   \n",
       "                 MEDIUM                      -1.451433   \n",
       "generalist_tv    EASY                        -6.609124   \n",
       "                 HARD                        -3.376755   \n",
       "                 MEDIUM                       0.869468   \n",
       "music_radio      EASY                        -2.263857   \n",
       "                 HARD                        71.434134   \n",
       "                 MEDIUM                       4.543460   \n",
       "news_tv          EASY                         0.435185   \n",
       "                 HARD                        -9.318224   \n",
       "                 MEDIUM                       1.590077   \n",
       "\n",
       "                                 (IER, identification error rate, %)  \\\n",
       "channel_category vad_difficulty                                        \n",
       "generalist_radio EASY                                       2.432442   \n",
       "                 HARD                                      35.483808   \n",
       "                 MEDIUM                                    11.638845   \n",
       "generalist_tv    EASY                                      15.989948   \n",
       "                 HARD                                      54.486051   \n",
       "                 MEDIUM                                    18.293169   \n",
       "music_radio      EASY                                       2.545830   \n",
       "                 HARD                                     652.633362   \n",
       "                 MEDIUM                                    10.021808   \n",
       "news_tv          EASY                                       3.082951   \n",
       "                 HARD                                      32.701088   \n",
       "                 MEDIUM                                    12.546062   \n",
       "\n",
       "                                 (IER Female, identification error rate, %)  \\\n",
       "channel_category vad_difficulty                                               \n",
       "generalist_radio EASY                                              1.714421   \n",
       "                 HARD                                             30.055415   \n",
       "                 MEDIUM                                           77.481359   \n",
       "generalist_tv    EASY                                             23.455468   \n",
       "                 HARD                                             48.741820   \n",
       "                 MEDIUM                                           17.229385   \n",
       "music_radio      EASY                                              2.347034   \n",
       "                 HARD                                             62.151487   \n",
       "                 MEDIUM                                            6.496364   \n",
       "news_tv          EASY                                              2.725143   \n",
       "                 HARD                                             28.590344   \n",
       "                 MEDIUM                                           13.098857   \n",
       "\n",
       "                                 (IER Male, identification error rate, %)  \\\n",
       "channel_category vad_difficulty                                             \n",
       "generalist_radio EASY                                            2.611223   \n",
       "                 HARD                                           37.800322   \n",
       "                 MEDIUM                                         10.105248   \n",
       "generalist_tv    EASY                                            8.103532   \n",
       "                 HARD                                           58.900198   \n",
       "                 MEDIUM                                         18.868623   \n",
       "music_radio      EASY                                            2.679484   \n",
       "                 HARD                                         2240.037477   \n",
       "                 MEDIUM                                         11.822060   \n",
       "news_tv          EASY                                            3.419553   \n",
       "                 HARD                                           35.058109   \n",
       "                 MEDIUM                                         12.290901   \n",
       "\n",
       "                                 nbfile  WSTP_RMSE  IER gender bias  \n",
       "channel_category vad_difficulty                                      \n",
       "generalist_radio EASY                25   3.025124        -0.896802  \n",
       "                 HARD                 3   9.663539        -7.744907  \n",
       "                 MEDIUM               6   5.160619        67.376111  \n",
       "generalist_tv    EASY                17  20.367249        15.351936  \n",
       "                 HARD                27  19.679952       -10.158377  \n",
       "                 MEDIUM              49  12.877909        -1.639238  \n",
       "music_radio      EASY                25   2.069831        -0.332450  \n",
       "                 HARD                13  29.516576     -2177.885990  \n",
       "                 MEDIUM               8   1.014231        -5.325696  \n",
       "news_tv          EASY                28   2.951893        -0.694410  \n",
       "                 HARD                 1   9.318224        -6.467765  \n",
       "                 MEDIUM              15   5.572590         0.807956  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# multi criterion categories\n",
    "details, categories = ge.compare_category(iss, ['channel_category', 'vad_difficulty'])\n",
    "simplify_results(categories)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "d3c0edc0",
   "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>(WSTP, WSTP error, %)</th>\n",
       "      <th>(IER, identification error rate, %)</th>\n",
       "      <th>(IER Female, identification error rate, %)</th>\n",
       "      <th>(IER Male, identification error rate, %)</th>\n",
       "      <th>nbfile</th>\n",
       "      <th>WSTP_RMSE</th>\n",
       "      <th>IER gender bias</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>channel_name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Arte</th>\n",
       "      <td>3.903907</td>\n",
       "      <td>26.765495</td>\n",
       "      <td>17.727759</td>\n",
       "      <td>29.174302</td>\n",
       "      <td>6</td>\n",
       "      <td>11.678924</td>\n",
       "      <td>-11.446543</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BFM TV</th>\n",
       "      <td>0.717261</td>\n",
       "      <td>5.463057</td>\n",
       "      <td>5.583742</td>\n",
       "      <td>5.390309</td>\n",
       "      <td>8</td>\n",
       "      <td>1.458577</td>\n",
       "      <td>0.193432</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNEWS</th>\n",
       "      <td>0.243643</td>\n",
       "      <td>4.704669</td>\n",
       "      <td>5.343231</td>\n",
       "      <td>4.346259</td>\n",
       "      <td>11</td>\n",
       "      <td>3.743300</td>\n",
       "      <td>0.996972</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Canal+</th>\n",
       "      <td>1.677911</td>\n",
       "      <td>42.855639</td>\n",
       "      <td>28.477517</td>\n",
       "      <td>52.482291</td>\n",
       "      <td>10</td>\n",
       "      <td>11.717505</td>\n",
       "      <td>-24.004775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Chérie 25</th>\n",
       "      <td>6.900205</td>\n",
       "      <td>6.250510</td>\n",
       "      <td>0.049898</td>\n",
       "      <td>21.353777</td>\n",
       "      <td>1</td>\n",
       "      <td>6.900205</td>\n",
       "      <td>-21.303879</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FIP</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France 2</th>\n",
       "      <td>2.478060</td>\n",
       "      <td>13.974871</td>\n",
       "      <td>13.293729</td>\n",
       "      <td>14.515860</td>\n",
       "      <td>8</td>\n",
       "      <td>10.289955</td>\n",
       "      <td>-1.222131</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France 24</th>\n",
       "      <td>0.425246</td>\n",
       "      <td>9.256023</td>\n",
       "      <td>9.472003</td>\n",
       "      <td>9.116537</td>\n",
       "      <td>11</td>\n",
       "      <td>3.710358</td>\n",
       "      <td>0.355467</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France 3</th>\n",
       "      <td>-4.280538</td>\n",
       "      <td>26.249833</td>\n",
       "      <td>49.209521</td>\n",
       "      <td>18.497738</td>\n",
       "      <td>8</td>\n",
       "      <td>15.862629</td>\n",
       "      <td>30.711783</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France 4</th>\n",
       "      <td>-21.490039</td>\n",
       "      <td>40.313481</td>\n",
       "      <td>157.508364</td>\n",
       "      <td>19.454062</td>\n",
       "      <td>12</td>\n",
       "      <td>35.627978</td>\n",
       "      <td>138.054302</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France 5</th>\n",
       "      <td>-4.004677</td>\n",
       "      <td>20.481783</td>\n",
       "      <td>37.732615</td>\n",
       "      <td>15.574525</td>\n",
       "      <td>5</td>\n",
       "      <td>13.130487</td>\n",
       "      <td>22.158090</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France Bleu</th>\n",
       "      <td>-3.293404</td>\n",
       "      <td>2.499290</td>\n",
       "      <td>2.347034</td>\n",
       "      <td>2.776031</td>\n",
       "      <td>12</td>\n",
       "      <td>2.987543</td>\n",
       "      <td>-0.428997</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France Culture</th>\n",
       "      <td>-1.808778</td>\n",
       "      <td>23.602622</td>\n",
       "      <td>45.096239</td>\n",
       "      <td>20.330828</td>\n",
       "      <td>6</td>\n",
       "      <td>8.525927</td>\n",
       "      <td>24.765411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France Info</th>\n",
       "      <td>0.363075</td>\n",
       "      <td>2.591640</td>\n",
       "      <td>2.060057</td>\n",
       "      <td>2.770883</td>\n",
       "      <td>11</td>\n",
       "      <td>0.747023</td>\n",
       "      <td>-0.710825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France Inter</th>\n",
       "      <td>0.616180</td>\n",
       "      <td>5.130897</td>\n",
       "      <td>1.796961</td>\n",
       "      <td>6.297948</td>\n",
       "      <td>6</td>\n",
       "      <td>0.892282</td>\n",
       "      <td>-4.500986</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France Musique</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.353755</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.353755</td>\n",
       "      <td>14</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-4.353755</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fun Radio</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Gulli</th>\n",
       "      <td>-11.610566</td>\n",
       "      <td>105.722171</td>\n",
       "      <td>115.040224</td>\n",
       "      <td>70.867859</td>\n",
       "      <td>1</td>\n",
       "      <td>11.610566</td>\n",
       "      <td>44.172365</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LCI</th>\n",
       "      <td>1.960561</td>\n",
       "      <td>6.529352</td>\n",
       "      <td>3.392661</td>\n",
       "      <td>10.240194</td>\n",
       "      <td>14</td>\n",
       "      <td>5.818869</td>\n",
       "      <td>-6.847533</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>M6</th>\n",
       "      <td>0.303646</td>\n",
       "      <td>14.245188</td>\n",
       "      <td>7.855940</td>\n",
       "      <td>25.273896</td>\n",
       "      <td>10</td>\n",
       "      <td>8.049367</td>\n",
       "      <td>-17.417957</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mouv'</th>\n",
       "      <td>37.674876</td>\n",
       "      <td>106.321413</td>\n",
       "      <td>14.193901</td>\n",
       "      <td>414.095545</td>\n",
       "      <td>12</td>\n",
       "      <td>30.732985</td>\n",
       "      <td>-399.901644</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NRJ 12</th>\n",
       "      <td>-0.329440</td>\n",
       "      <td>26.841334</td>\n",
       "      <td>15.781700</td>\n",
       "      <td>46.714764</td>\n",
       "      <td>3</td>\n",
       "      <td>11.477451</td>\n",
       "      <td>-30.933064</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Paris Première</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>9.054200</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>9.054200</td>\n",
       "      <td>1</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-9.054200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RMC</th>\n",
       "      <td>2.274843</td>\n",
       "      <td>2.293162</td>\n",
       "      <td>0.503868</td>\n",
       "      <td>2.473301</td>\n",
       "      <td>9</td>\n",
       "      <td>4.812603</td>\n",
       "      <td>-1.969433</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RTL</th>\n",
       "      <td>1.862377</td>\n",
       "      <td>3.348031</td>\n",
       "      <td>0.280505</td>\n",
       "      <td>3.439349</td>\n",
       "      <td>2</td>\n",
       "      <td>2.570399</td>\n",
       "      <td>-3.158845</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Skyrock</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TF1</th>\n",
       "      <td>-0.825360</td>\n",
       "      <td>30.163099</td>\n",
       "      <td>35.248752</td>\n",
       "      <td>28.169427</td>\n",
       "      <td>13</td>\n",
       "      <td>10.458249</td>\n",
       "      <td>7.079325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TFX</th>\n",
       "      <td>26.248015</td>\n",
       "      <td>84.644264</td>\n",
       "      <td>46.626476</td>\n",
       "      <td>195.868567</td>\n",
       "      <td>2</td>\n",
       "      <td>18.022931</td>\n",
       "      <td>-149.242091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TV5 Monde</th>\n",
       "      <td>2.239385</td>\n",
       "      <td>16.188088</td>\n",
       "      <td>13.939947</td>\n",
       "      <td>19.910299</td>\n",
       "      <td>13</td>\n",
       "      <td>9.577796</td>\n",
       "      <td>-5.970352</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                (WSTP, WSTP error, %)  (IER, identification error rate, %)  \\\n",
       "channel_name                                                                 \n",
       "Arte                         3.903907                            26.765495   \n",
       "BFM TV                       0.717261                             5.463057   \n",
       "CNEWS                        0.243643                             4.704669   \n",
       "Canal+                       1.677911                            42.855639   \n",
       "Chérie 25                    6.900205                             6.250510   \n",
       "FIP                          0.000000                             0.000000   \n",
       "France 2                     2.478060                            13.974871   \n",
       "France 24                    0.425246                             9.256023   \n",
       "France 3                    -4.280538                            26.249833   \n",
       "France 4                   -21.490039                            40.313481   \n",
       "France 5                    -4.004677                            20.481783   \n",
       "France Bleu                 -3.293404                             2.499290   \n",
       "France Culture              -1.808778                            23.602622   \n",
       "France Info                  0.363075                             2.591640   \n",
       "France Inter                 0.616180                             5.130897   \n",
       "France Musique               0.000000                             4.353755   \n",
       "Fun Radio                    0.000000                             0.000000   \n",
       "Gulli                      -11.610566                           105.722171   \n",
       "LCI                          1.960561                             6.529352   \n",
       "M6                           0.303646                            14.245188   \n",
       "Mouv'                       37.674876                           106.321413   \n",
       "NRJ 12                      -0.329440                            26.841334   \n",
       "Paris Première               0.000000                             9.054200   \n",
       "RMC                          2.274843                             2.293162   \n",
       "RTL                          1.862377                             3.348031   \n",
       "Skyrock                      0.000000                             0.000000   \n",
       "TF1                         -0.825360                            30.163099   \n",
       "TFX                         26.248015                            84.644264   \n",
       "TV5 Monde                    2.239385                            16.188088   \n",
       "\n",
       "                (IER Female, identification error rate, %)  \\\n",
       "channel_name                                                 \n",
       "Arte                                             17.727759   \n",
       "BFM TV                                            5.583742   \n",
       "CNEWS                                             5.343231   \n",
       "Canal+                                           28.477517   \n",
       "Chérie 25                                         0.049898   \n",
       "FIP                                               0.000000   \n",
       "France 2                                         13.293729   \n",
       "France 24                                         9.472003   \n",
       "France 3                                         49.209521   \n",
       "France 4                                        157.508364   \n",
       "France 5                                         37.732615   \n",
       "France Bleu                                       2.347034   \n",
       "France Culture                                   45.096239   \n",
       "France Info                                       2.060057   \n",
       "France Inter                                      1.796961   \n",
       "France Musique                                    0.000000   \n",
       "Fun Radio                                         0.000000   \n",
       "Gulli                                           115.040224   \n",
       "LCI                                               3.392661   \n",
       "M6                                                7.855940   \n",
       "Mouv'                                            14.193901   \n",
       "NRJ 12                                           15.781700   \n",
       "Paris Première                                    0.000000   \n",
       "RMC                                               0.503868   \n",
       "RTL                                               0.280505   \n",
       "Skyrock                                           0.000000   \n",
       "TF1                                              35.248752   \n",
       "TFX                                              46.626476   \n",
       "TV5 Monde                                        13.939947   \n",
       "\n",
       "                (IER Male, identification error rate, %)  nbfile  WSTP_RMSE  \\\n",
       "channel_name                                                                  \n",
       "Arte                                           29.174302       6  11.678924   \n",
       "BFM TV                                          5.390309       8   1.458577   \n",
       "CNEWS                                           4.346259      11   3.743300   \n",
       "Canal+                                         52.482291      10  11.717505   \n",
       "Chérie 25                                      21.353777       1   6.900205   \n",
       "FIP                                             0.000000       4   0.000000   \n",
       "France 2                                       14.515860       8  10.289955   \n",
       "France 24                                       9.116537      11   3.710358   \n",
       "France 3                                       18.497738       8  15.862629   \n",
       "France 4                                       19.454062      12  35.627978   \n",
       "France 5                                       15.574525       5  13.130487   \n",
       "France Bleu                                     2.776031      12   2.987543   \n",
       "France Culture                                 20.330828       6   8.525927   \n",
       "France Info                                     2.770883      11   0.747023   \n",
       "France Inter                                    6.297948       6   0.892282   \n",
       "France Musique                                  4.353755      14   0.000000   \n",
       "Fun Radio                                       0.000000       2   0.000000   \n",
       "Gulli                                          70.867859       1  11.610566   \n",
       "LCI                                            10.240194      14   5.818869   \n",
       "M6                                             25.273896      10   8.049367   \n",
       "Mouv'                                         414.095545      12  30.732985   \n",
       "NRJ 12                                         46.714764       3  11.477451   \n",
       "Paris Première                                  9.054200       1   0.000000   \n",
       "RMC                                             2.473301       9   4.812603   \n",
       "RTL                                             3.439349       2   2.570399   \n",
       "Skyrock                                         0.000000       2   0.000000   \n",
       "TF1                                            28.169427      13  10.458249   \n",
       "TFX                                           195.868567       2  18.022931   \n",
       "TV5 Monde                                      19.910299      13   9.577796   \n",
       "\n",
       "                IER gender bias  \n",
       "channel_name                     \n",
       "Arte                 -11.446543  \n",
       "BFM TV                 0.193432  \n",
       "CNEWS                  0.996972  \n",
       "Canal+               -24.004775  \n",
       "Chérie 25            -21.303879  \n",
       "FIP                    0.000000  \n",
       "France 2              -1.222131  \n",
       "France 24              0.355467  \n",
       "France 3              30.711783  \n",
       "France 4             138.054302  \n",
       "France 5              22.158090  \n",
       "France Bleu           -0.428997  \n",
       "France Culture        24.765411  \n",
       "France Info           -0.710825  \n",
       "France Inter          -4.500986  \n",
       "France Musique        -4.353755  \n",
       "Fun Radio              0.000000  \n",
       "Gulli                 44.172365  \n",
       "LCI                   -6.847533  \n",
       "M6                   -17.417957  \n",
       "Mouv'               -399.901644  \n",
       "NRJ 12               -30.933064  \n",
       "Paris Première        -9.054200  \n",
       "RMC                   -1.969433  \n",
       "RTL                   -3.158845  \n",
       "Skyrock                0.000000  \n",
       "TF1                    7.079325  \n",
       "TFX                 -149.242091  \n",
       "TV5 Monde             -5.970352  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# results per channel\n",
    "details, categories = ge.compare_category(iss, ['channel_name'])\n",
    "simplify_results(categories)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "5d68821d",
   "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>(WSTP, WSTP error, %)</th>\n",
       "      <th>(WSTP, ref_male_dur, )</th>\n",
       "      <th>(WSTP, ref_female_dur, )</th>\n",
       "      <th>(WSTP, hyp_male_dur, )</th>\n",
       "      <th>(WSTP, hyp_female_dur, )</th>\n",
       "      <th>(IER, identification error rate, %)</th>\n",
       "      <th>(IER, total, )</th>\n",
       "      <th>(IER, correct, )</th>\n",
       "      <th>(IER, correct, %)</th>\n",
       "      <th>(IER, false alarm, )</th>\n",
       "      <th>...</th>\n",
       "      <th>(IER Male, identification error rate, %)</th>\n",
       "      <th>(IER Male, total, )</th>\n",
       "      <th>(IER Male, correct, )</th>\n",
       "      <th>(IER Male, correct, %)</th>\n",
       "      <th>(IER Male, false alarm, )</th>\n",
       "      <th>(IER Male, false alarm, %)</th>\n",
       "      <th>(IER Male, missed detection, )</th>\n",
       "      <th>(IER Male, missed detection, %)</th>\n",
       "      <th>(IER Male, confusion, )</th>\n",
       "      <th>(IER Male, confusion, %)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>item</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>radio-FBL-011904</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>radio-FBL-012717</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>radio-FBL-025455</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>radio-FBL-035452</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>radio-FBL-044522</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tv-TF1-215118</th>\n",
       "      <td>-4.688474</td>\n",
       "      <td>5.724</td>\n",
       "      <td>13.731</td>\n",
       "      <td>5.10</td>\n",
       "      <td>15.52</td>\n",
       "      <td>10.443075</td>\n",
       "      <td>16.882</td>\n",
       "      <td>15.965</td>\n",
       "      <td>94.568179</td>\n",
       "      <td>0.846</td>\n",
       "      <td>...</td>\n",
       "      <td>2.263856</td>\n",
       "      <td>5.124</td>\n",
       "      <td>4.516</td>\n",
       "      <td>88.134270</td>\n",
       "      <td>0.116</td>\n",
       "      <td>2.263856</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tv-TF1-225026</th>\n",
       "      <td>-5.115000</td>\n",
       "      <td>37.768</td>\n",
       "      <td>8.924</td>\n",
       "      <td>46.10</td>\n",
       "      <td>14.74</td>\n",
       "      <td>32.358499</td>\n",
       "      <td>43.392</td>\n",
       "      <td>40.673</td>\n",
       "      <td>93.733868</td>\n",
       "      <td>11.322</td>\n",
       "      <td>...</td>\n",
       "      <td>25.079845</td>\n",
       "      <td>35.068</td>\n",
       "      <td>32.379</td>\n",
       "      <td>92.332041</td>\n",
       "      <td>8.765</td>\n",
       "      <td>24.994297</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.030</td>\n",
       "      <td>0.085548</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tv-TFX-230837</th>\n",
       "      <td>25.488273</td>\n",
       "      <td>2.828</td>\n",
       "      <td>31.343</td>\n",
       "      <td>13.58</td>\n",
       "      <td>26.64</td>\n",
       "      <td>81.186036</td>\n",
       "      <td>24.721</td>\n",
       "      <td>16.571</td>\n",
       "      <td>67.032078</td>\n",
       "      <td>11.920</td>\n",
       "      <td>...</td>\n",
       "      <td>1745.129225</td>\n",
       "      <td>0.503</td>\n",
       "      <td>0.460</td>\n",
       "      <td>91.451292</td>\n",
       "      <td>8.304</td>\n",
       "      <td>1650.894632</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.474</td>\n",
       "      <td>94.234592</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tv-TFX-231251</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>10.775</td>\n",
       "      <td>0.000</td>\n",
       "      <td>21.04</td>\n",
       "      <td>0.00</td>\n",
       "      <td>95.639871</td>\n",
       "      <td>7.775</td>\n",
       "      <td>7.775</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>7.436</td>\n",
       "      <td>...</td>\n",
       "      <td>95.639871</td>\n",
       "      <td>7.775</td>\n",
       "      <td>7.775</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>7.436</td>\n",
       "      <td>95.639871</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TOTAL</th>\n",
       "      <td>-0.227377</td>\n",
       "      <td>6482.576</td>\n",
       "      <td>3620.469</td>\n",
       "      <td>7121.76</td>\n",
       "      <td>4016.92</td>\n",
       "      <td>13.725407</td>\n",
       "      <td>8989.832</td>\n",
       "      <td>8529.305</td>\n",
       "      <td>94.877246</td>\n",
       "      <td>773.364</td>\n",
       "      <td>...</td>\n",
       "      <td>13.510989</td>\n",
       "      <td>5822.712</td>\n",
       "      <td>5537.030</td>\n",
       "      <td>95.093661</td>\n",
       "      <td>542.106</td>\n",
       "      <td>9.310198</td>\n",
       "      <td>149.812</td>\n",
       "      <td>2.57289</td>\n",
       "      <td>94.788</td>\n",
       "      <td>1.627901</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>278 rows × 35 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                  (WSTP, WSTP error, %)  (WSTP, ref_male_dur, )  \\\n",
       "item                                                              \n",
       "radio-FBL-011904               0.000000                   0.000   \n",
       "radio-FBL-012717               0.000000                   0.000   \n",
       "radio-FBL-025455               0.000000                   0.000   \n",
       "radio-FBL-035452               0.000000                   0.000   \n",
       "radio-FBL-044522               0.000000                   0.000   \n",
       "...                                 ...                     ...   \n",
       "tv-TF1-215118                 -4.688474                   5.724   \n",
       "tv-TF1-225026                 -5.115000                  37.768   \n",
       "tv-TFX-230837                 25.488273                   2.828   \n",
       "tv-TFX-231251                  0.000000                  10.775   \n",
       "TOTAL                         -0.227377                6482.576   \n",
       "\n",
       "                  (WSTP, ref_female_dur, )  (WSTP, hyp_male_dur, )  \\\n",
       "item                                                                 \n",
       "radio-FBL-011904                     0.000                    0.00   \n",
       "radio-FBL-012717                     0.000                    0.00   \n",
       "radio-FBL-025455                     0.000                    0.00   \n",
       "radio-FBL-035452                     0.000                    0.00   \n",
       "radio-FBL-044522                     0.000                    0.00   \n",
       "...                                    ...                     ...   \n",
       "tv-TF1-215118                       13.731                    5.10   \n",
       "tv-TF1-225026                        8.924                   46.10   \n",
       "tv-TFX-230837                       31.343                   13.58   \n",
       "tv-TFX-231251                        0.000                   21.04   \n",
       "TOTAL                             3620.469                 7121.76   \n",
       "\n",
       "                  (WSTP, hyp_female_dur, )  \\\n",
       "item                                         \n",
       "radio-FBL-011904                      0.00   \n",
       "radio-FBL-012717                      0.00   \n",
       "radio-FBL-025455                      0.00   \n",
       "radio-FBL-035452                      0.00   \n",
       "radio-FBL-044522                      0.00   \n",
       "...                                    ...   \n",
       "tv-TF1-215118                        15.52   \n",
       "tv-TF1-225026                        14.74   \n",
       "tv-TFX-230837                        26.64   \n",
       "tv-TFX-231251                         0.00   \n",
       "TOTAL                              4016.92   \n",
       "\n",
       "                  (IER, identification error rate, %)  (IER, total, )  \\\n",
       "item                                                                    \n",
       "radio-FBL-011904                             0.000000           0.000   \n",
       "radio-FBL-012717                             0.000000           0.000   \n",
       "radio-FBL-025455                             0.000000           0.000   \n",
       "radio-FBL-035452                             0.000000           0.000   \n",
       "radio-FBL-044522                             0.000000           0.000   \n",
       "...                                               ...             ...   \n",
       "tv-TF1-215118                               10.443075          16.882   \n",
       "tv-TF1-225026                               32.358499          43.392   \n",
       "tv-TFX-230837                               81.186036          24.721   \n",
       "tv-TFX-231251                               95.639871           7.775   \n",
       "TOTAL                                       13.725407        8989.832   \n",
       "\n",
       "                  (IER, correct, )  (IER, correct, %)  (IER, false alarm, )  \\\n",
       "item                                                                          \n",
       "radio-FBL-011904             0.000                NaN                 0.000   \n",
       "radio-FBL-012717             0.000                NaN                 0.000   \n",
       "radio-FBL-025455             0.000                NaN                 0.000   \n",
       "radio-FBL-035452             0.000                NaN                 0.000   \n",
       "radio-FBL-044522             0.000                NaN                 0.000   \n",
       "...                            ...                ...                   ...   \n",
       "tv-TF1-215118               15.965          94.568179                 0.846   \n",
       "tv-TF1-225026               40.673          93.733868                11.322   \n",
       "tv-TFX-230837               16.571          67.032078                11.920   \n",
       "tv-TFX-231251                7.775         100.000000                 7.436   \n",
       "TOTAL                     8529.305          94.877246               773.364   \n",
       "\n",
       "                  ...  (IER Male, identification error rate, %)  \\\n",
       "item              ...                                             \n",
       "radio-FBL-011904  ...                                  0.000000   \n",
       "radio-FBL-012717  ...                                  0.000000   \n",
       "radio-FBL-025455  ...                                  0.000000   \n",
       "radio-FBL-035452  ...                                  0.000000   \n",
       "radio-FBL-044522  ...                                  0.000000   \n",
       "...               ...                                       ...   \n",
       "tv-TF1-215118     ...                                  2.263856   \n",
       "tv-TF1-225026     ...                                 25.079845   \n",
       "tv-TFX-230837     ...                               1745.129225   \n",
       "tv-TFX-231251     ...                                 95.639871   \n",
       "TOTAL             ...                                 13.510989   \n",
       "\n",
       "                  (IER Male, total, )  (IER Male, correct, )  \\\n",
       "item                                                           \n",
       "radio-FBL-011904                0.000                  0.000   \n",
       "radio-FBL-012717                0.000                  0.000   \n",
       "radio-FBL-025455                0.000                  0.000   \n",
       "radio-FBL-035452                0.000                  0.000   \n",
       "radio-FBL-044522                0.000                  0.000   \n",
       "...                               ...                    ...   \n",
       "tv-TF1-215118                   5.124                  4.516   \n",
       "tv-TF1-225026                  35.068                 32.379   \n",
       "tv-TFX-230837                   0.503                  0.460   \n",
       "tv-TFX-231251                   7.775                  7.775   \n",
       "TOTAL                        5822.712               5537.030   \n",
       "\n",
       "                  (IER Male, correct, %)  (IER Male, false alarm, )  \\\n",
       "item                                                                  \n",
       "radio-FBL-011904                     NaN                      0.000   \n",
       "radio-FBL-012717                     NaN                      0.000   \n",
       "radio-FBL-025455                     NaN                      0.000   \n",
       "radio-FBL-035452                     NaN                      0.000   \n",
       "radio-FBL-044522                     NaN                      0.000   \n",
       "...                                  ...                        ...   \n",
       "tv-TF1-215118                  88.134270                      0.116   \n",
       "tv-TF1-225026                  92.332041                      8.765   \n",
       "tv-TFX-230837                  91.451292                      8.304   \n",
       "tv-TFX-231251                 100.000000                      7.436   \n",
       "TOTAL                          95.093661                    542.106   \n",
       "\n",
       "                  (IER Male, false alarm, %)  (IER Male, missed detection, )  \\\n",
       "item                                                                           \n",
       "radio-FBL-011904                         NaN                           0.000   \n",
       "radio-FBL-012717                         NaN                           0.000   \n",
       "radio-FBL-025455                         NaN                           0.000   \n",
       "radio-FBL-035452                         NaN                           0.000   \n",
       "radio-FBL-044522                         NaN                           0.000   \n",
       "...                                      ...                             ...   \n",
       "tv-TF1-215118                       2.263856                           0.000   \n",
       "tv-TF1-225026                      24.994297                           0.000   \n",
       "tv-TFX-230837                    1650.894632                           0.000   \n",
       "tv-TFX-231251                      95.639871                           0.000   \n",
       "TOTAL                               9.310198                         149.812   \n",
       "\n",
       "                  (IER Male, missed detection, %)  (IER Male, confusion, )  \\\n",
       "item                                                                         \n",
       "radio-FBL-011904                              NaN                    0.000   \n",
       "radio-FBL-012717                              NaN                    0.000   \n",
       "radio-FBL-025455                              NaN                    0.000   \n",
       "radio-FBL-035452                              NaN                    0.000   \n",
       "radio-FBL-044522                              NaN                    0.000   \n",
       "...                                           ...                      ...   \n",
       "tv-TF1-215118                             0.00000                    0.000   \n",
       "tv-TF1-225026                             0.00000                    0.030   \n",
       "tv-TFX-230837                             0.00000                    0.474   \n",
       "tv-TFX-231251                             0.00000                    0.000   \n",
       "TOTAL                                     2.57289                   94.788   \n",
       "\n",
       "                  (IER Male, confusion, %)  \n",
       "item                                        \n",
       "radio-FBL-011904                       NaN  \n",
       "radio-FBL-012717                       NaN  \n",
       "radio-FBL-025455                       NaN  \n",
       "radio-FBL-035452                       NaN  \n",
       "radio-FBL-044522                       NaN  \n",
       "...                                    ...  \n",
       "tv-TF1-215118                     0.000000  \n",
       "tv-TF1-225026                     0.085548  \n",
       "tv-TFX-230837                    94.234592  \n",
       "tv-TFX-231251                     0.000000  \n",
       "TOTAL                             1.627901  \n",
       "\n",
       "[278 rows x 35 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# possibilité de comparer des listes de fichier\n",
    "# la derniere ligne correspond aux stats globales\n",
    "results = ge.compare_lfiles(lref, liss)\n",
    "results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "ec80efaf",
   "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>(WSTP, WSTP error, %)</th>\n",
       "      <th>(IER, identification error rate, %)</th>\n",
       "      <th>(IER Female, identification error rate, %)</th>\n",
       "      <th>(IER Male, identification error rate, %)</th>\n",
       "      <th>IER gender bias</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>item</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>radio-FBL-011904</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>radio-FBL-012717</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>radio-FBL-025455</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>radio-FBL-035452</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>radio-FBL-044522</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tv-TF1-215118</th>\n",
       "      <td>-4.688474</td>\n",
       "      <td>10.443075</td>\n",
       "      <td>14.007484</td>\n",
       "      <td>2.263856</td>\n",
       "      <td>11.743628</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tv-TF1-225026</th>\n",
       "      <td>-5.115000</td>\n",
       "      <td>32.358499</td>\n",
       "      <td>63.022585</td>\n",
       "      <td>25.079845</td>\n",
       "      <td>37.942740</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tv-TFX-230837</th>\n",
       "      <td>25.488273</td>\n",
       "      <td>81.186036</td>\n",
       "      <td>46.626476</td>\n",
       "      <td>1745.129225</td>\n",
       "      <td>-1698.502748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tv-TFX-231251</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>95.639871</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>95.639871</td>\n",
       "      <td>-95.639871</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TOTAL</th>\n",
       "      <td>-0.227377</td>\n",
       "      <td>13.725407</td>\n",
       "      <td>14.119610</td>\n",
       "      <td>13.510989</td>\n",
       "      <td>0.608621</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>278 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                  (WSTP, WSTP error, %)  (IER, identification error rate, %)  \\\n",
       "item                                                                           \n",
       "radio-FBL-011904               0.000000                             0.000000   \n",
       "radio-FBL-012717               0.000000                             0.000000   \n",
       "radio-FBL-025455               0.000000                             0.000000   \n",
       "radio-FBL-035452               0.000000                             0.000000   \n",
       "radio-FBL-044522               0.000000                             0.000000   \n",
       "...                                 ...                                  ...   \n",
       "tv-TF1-215118                 -4.688474                            10.443075   \n",
       "tv-TF1-225026                 -5.115000                            32.358499   \n",
       "tv-TFX-230837                 25.488273                            81.186036   \n",
       "tv-TFX-231251                  0.000000                            95.639871   \n",
       "TOTAL                         -0.227377                            13.725407   \n",
       "\n",
       "                  (IER Female, identification error rate, %)  \\\n",
       "item                                                           \n",
       "radio-FBL-011904                                    0.000000   \n",
       "radio-FBL-012717                                    0.000000   \n",
       "radio-FBL-025455                                    0.000000   \n",
       "radio-FBL-035452                                    0.000000   \n",
       "radio-FBL-044522                                    0.000000   \n",
       "...                                                      ...   \n",
       "tv-TF1-215118                                      14.007484   \n",
       "tv-TF1-225026                                      63.022585   \n",
       "tv-TFX-230837                                      46.626476   \n",
       "tv-TFX-231251                                       0.000000   \n",
       "TOTAL                                              14.119610   \n",
       "\n",
       "                  (IER Male, identification error rate, %)  IER gender bias  \n",
       "item                                                                         \n",
       "radio-FBL-011904                                  0.000000         0.000000  \n",
       "radio-FBL-012717                                  0.000000         0.000000  \n",
       "radio-FBL-025455                                  0.000000         0.000000  \n",
       "radio-FBL-035452                                  0.000000         0.000000  \n",
       "radio-FBL-044522                                  0.000000         0.000000  \n",
       "...                                                    ...              ...  \n",
       "tv-TF1-215118                                     2.263856        11.743628  \n",
       "tv-TF1-225026                                    25.079845        37.942740  \n",
       "tv-TFX-230837                                  1745.129225     -1698.502748  \n",
       "tv-TFX-231251                                    95.639871       -95.639871  \n",
       "TOTAL                                            13.510989         0.608621  \n",
       "\n",
       "[278 rows x 5 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "simplify_results(results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "72235eec",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>(WSTP, WSTP error, %)</th>\n",
       "      <th>(IER, identification error rate, %)</th>\n",
       "      <th>(IER Female, identification error rate, %)</th>\n",
       "      <th>(IER Male, identification error rate, %)</th>\n",
       "      <th>nbfile</th>\n",
       "      <th>WSTP_RMSE</th>\n",
       "      <th>IER gender bias</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>channel_category</th>\n",
       "      <th>broadcast_hour</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">generalist_radio</th>\n",
       "      <th>0</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.880724</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.880724</td>\n",
       "      <td>1</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-3.880724</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>21.430471</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>21.430471</td>\n",
       "      <td>1</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-21.430471</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5.501004</td>\n",
       "      <td>10.540243</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>11.106713</td>\n",
       "      <td>3</td>\n",
       "      <td>7.992228</td>\n",
       "      <td>-11.106713</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.995532</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.995532</td>\n",
       "      <td>1</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-1.995532</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>8.138296</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>8.138296</td>\n",
       "      <td>1</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-8.138296</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">news_tv</th>\n",
       "      <th>19</th>\n",
       "      <td>-0.242937</td>\n",
       "      <td>1.878058</td>\n",
       "      <td>9.350057</td>\n",
       "      <td>1.603370</td>\n",
       "      <td>1</td>\n",
       "      <td>0.242937</td>\n",
       "      <td>7.746687</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.628497</td>\n",
       "      <td>11.602302</td>\n",
       "      <td>6.141773</td>\n",
       "      <td>16.686472</td>\n",
       "      <td>3</td>\n",
       "      <td>5.578046</td>\n",
       "      <td>-10.544699</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.847608</td>\n",
       "      <td>1.756323</td>\n",
       "      <td>1.184746</td>\n",
       "      <td>3.384086</td>\n",
       "      <td>2</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-2.199340</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.899579</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.899579</td>\n",
       "      <td>1</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-3.899579</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.160509</td>\n",
       "      <td>5.226709</td>\n",
       "      <td>0.817439</td>\n",
       "      <td>5.304439</td>\n",
       "      <td>3</td>\n",
       "      <td>0.141018</td>\n",
       "      <td>-4.487000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>83 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 (WSTP, WSTP error, %)  \\\n",
       "channel_category broadcast_hour                          \n",
       "generalist_radio 0                            0.000000   \n",
       "                 2                            0.000000   \n",
       "                 3                            5.501004   \n",
       "                 4                            0.000000   \n",
       "                 8                            0.000000   \n",
       "...                                                ...   \n",
       "news_tv          19                          -0.242937   \n",
       "                 20                           0.628497   \n",
       "                 21                           0.847608   \n",
       "                 22                           0.000000   \n",
       "                 23                           0.160509   \n",
       "\n",
       "                                 (IER, identification error rate, %)  \\\n",
       "channel_category broadcast_hour                                        \n",
       "generalist_radio 0                                          3.880724   \n",
       "                 2                                         21.430471   \n",
       "                 3                                         10.540243   \n",
       "                 4                                          1.995532   \n",
       "                 8                                          8.138296   \n",
       "...                                                              ...   \n",
       "news_tv          19                                         1.878058   \n",
       "                 20                                        11.602302   \n",
       "                 21                                         1.756323   \n",
       "                 22                                         3.899579   \n",
       "                 23                                         5.226709   \n",
       "\n",
       "                                 (IER Female, identification error rate, %)  \\\n",
       "channel_category broadcast_hour                                               \n",
       "generalist_radio 0                                                 0.000000   \n",
       "                 2                                                 0.000000   \n",
       "                 3                                                 0.000000   \n",
       "                 4                                                 0.000000   \n",
       "                 8                                                 0.000000   \n",
       "...                                                                     ...   \n",
       "news_tv          19                                                9.350057   \n",
       "                 20                                                6.141773   \n",
       "                 21                                                1.184746   \n",
       "                 22                                                0.000000   \n",
       "                 23                                                0.817439   \n",
       "\n",
       "                                 (IER Male, identification error rate, %)  \\\n",
       "channel_category broadcast_hour                                             \n",
       "generalist_radio 0                                               3.880724   \n",
       "                 2                                              21.430471   \n",
       "                 3                                              11.106713   \n",
       "                 4                                               1.995532   \n",
       "                 8                                               8.138296   \n",
       "...                                                                   ...   \n",
       "news_tv          19                                              1.603370   \n",
       "                 20                                             16.686472   \n",
       "                 21                                              3.384086   \n",
       "                 22                                              3.899579   \n",
       "                 23                                              5.304439   \n",
       "\n",
       "                                 nbfile  WSTP_RMSE  IER gender bias  \n",
       "channel_category broadcast_hour                                      \n",
       "generalist_radio 0                    1   0.000000        -3.880724  \n",
       "                 2                    1   0.000000       -21.430471  \n",
       "                 3                    3   7.992228       -11.106713  \n",
       "                 4                    1   0.000000        -1.995532  \n",
       "                 8                    1   0.000000        -8.138296  \n",
       "...                                 ...        ...              ...  \n",
       "news_tv          19                   1   0.242937         7.746687  \n",
       "                 20                   3   5.578046       -10.544699  \n",
       "                 21                   2   0.000000        -2.199340  \n",
       "                 22                   1   0.000000        -3.899579  \n",
       "                 23                   3   0.141018        -4.487000  \n",
       "\n",
       "[83 rows x 7 columns]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "details, categories = ge.compare_category(iss, ['channel_category', 'broadcast_hour'])\n",
    "sr = simplify_results(categories)\n",
    "sr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "083f95ce",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "generalist_radio\n",
      "generalist_tv\n",
      "music_radio\n",
      "news_tv\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7efbbcf248e0>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pylab as pl\n",
    "for k, sdf in sr.groupby('channel_category'):\n",
    "    print(k)\n",
    "    x = sdf.index.get_level_values(1)\n",
    "    y = sdf.WSTP_RMSE\n",
    "    pl.plot(x, y, label=k)\n",
    "pl.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "2b71af43",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "generalist_radio\n",
      "generalist_tv\n",
      "music_radio\n",
      "news_tv\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7efbbcedceb0>"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pylab as pl\n",
    "for k, sdf in sr.groupby('channel_category'):\n",
    "    print(k)\n",
    "    x = sdf.index.get_level_values(1)\n",
    "    y = sdf[('IER', 'identification error rate', '%')]\n",
    "    y = [min(e, 100) for e in y]\n",
    "    pl.plot(x, y, label=k)\n",
    "pl.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4850b321",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}