notebook/wer_calculation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<font size=\"5\"> Amazon Transcribe Accuracy Evaluation Using WER </font>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Install jiwer package"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: jiwer in /Users/houruize/opt/anaconda3/lib/python3.8/site-packages (2.2.0)\r\n",
      "Requirement already satisfied: python-Levenshtein in /Users/houruize/opt/anaconda3/lib/python3.8/site-packages (from jiwer) (0.12.2)\r\n",
      "Requirement already satisfied: numpy in /Users/houruize/opt/anaconda3/lib/python3.8/site-packages (from jiwer) (1.19.2)\r\n",
      "Requirement already satisfied: setuptools in /Users/houruize/opt/anaconda3/lib/python3.8/site-packages (from python-Levenshtein->jiwer) (50.3.1.post20201107)\r\n"
     ]
    }
   ],
   "source": [
    "!pip install jiwer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Import wer() function that calculates the word error rate from the jiwer package,\n",
    "also import other packages that are required in doing data analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "from jiwer import wer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Read hypothesis text file generated by Amazon Transcribe for different accent"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since this notebook is in the same directory with all the files needed, we read those files directly. All the audio and text files are uploaded in audio-text folder, when you read the file, make sure to add corresponding path to the file name."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('accent text.txt', 'r') as reader:\n",
    "    test = reader.readlines()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create a dataframe df that has \"Text Type\", \"Accent\", and \"Speed\" for each hypothesis text as variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "wer_data = {'Text Type':[1,2,3,4,5,1,2,3,4,5,1,2,3,4,5,1,2,3,4,5,1,2,3,4,5,1,2,3,4,5,\n",
    "                        1,2,3,4,5,1,2,3,4,5,1,2,3,4,5,1,2,3,4,5,1,2,3,4,5],\n",
    "            'Accent':['French', 'French', 'French', 'French', 'French','Japanese','Japanese','Japanese','Japanese','Japanese','Korean','Korean','Korean','Korean','Korean','Spanish','Spanish','Spanish','Spanish','Spanish','UKS','UKS','UKS','UKS','UKS','USS','USS','USS','USS','USS',\n",
    "                     'USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS','USS'],\n",
    "            'Speed':[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,\n",
    "                     0.5,0.5,0.5,0.5,0.5,0.8,0.8,0.8,0.8,0.8,1.0,1.0,1.0,1.0,1.0,1.2,1.2,1.2,1.2,1.2,1.5,1.5,1.5,1.5,1.5]}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "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>Text Type</th>\n",
       "      <th>Accent</th>\n",
       "      <th>Speed</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>French</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>French</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>French</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>French</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>French</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Text Type  Accent  Speed\n",
       "0          1  French    1.0\n",
       "1          2  French    1.0\n",
       "2          3  French    1.0\n",
       "3          4  French    1.0\n",
       "4          5  French    1.0"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.DataFrame(wer_data)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Keep the first 30 entries of df that are data for different accents, store it in df_accent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_accent = df[:30]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Read the five reference texts, save them into five variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"emoryqss.txt\", \"r\") as reader:\n",
    "    origin1 = reader.readlines()\n",
    "with open(\"emory1.txt\", \"r\") as reader:\n",
    "    origin2 = reader.readlines()\n",
    "with open(\"gatsby.txt\", \"r\") as reader:\n",
    "    origin3 = reader.readlines()\n",
    "with open(\"aws.txt\", \"r\") as reader:\n",
    "    origin4 = reader.readlines()\n",
    "with open(\"biden.txt\", \"r\") as reader:\n",
    "    origin5 = reader.readlines()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Calculate the WER for each hypothesis text, the result is stored in a list called accent_wer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "origin_ls = [origin1, origin2, origin3, origin4, origin5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "len_origin = len(origin_ls)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "origin = origin_ls*int((len(df_accent)/len_origin))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "accent_wer = []\n",
    "\n",
    "for i in range(len(df_accent)):\n",
    "    accent_wer.append(wer(origin[i], test[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.6336633663366337, 0.48148148148148145, 0.6666666666666666, 0.4262295081967213, 0.45794392523364486, 0.7425742574257426, 0.8296296296296296, 0.803921568627451, 0.6721311475409836, 0.8037383177570093, 0.801980198019802, 0.725925925925926, 0.8823529411764706, 0.5737704918032787, 0.616822429906542, 0.7326732673267327, 0.4148148148148148, 0.45098039215686275, 0.2786885245901639, 0.42990654205607476, 0.26732673267326734, 0.14814814814814814, 0.2549019607843137, 0.14754098360655737, 0.07476635514018691, 0.3069306930693069, 0.14814814814814814, 0.2549019607843137, 0.06557377049180328, 0.07476635514018691]\n"
     ]
    }
   ],
   "source": [
    "print(accent_wer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Add WER values stored in accent_wer to df_accent as column \"wer\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    Text Type    Accent  Speed       wer\n",
      "0           1    French    1.0  0.633663\n",
      "1           2    French    1.0  0.481481\n",
      "2           3    French    1.0  0.666667\n",
      "3           4    French    1.0  0.426230\n",
      "4           5    French    1.0  0.457944\n",
      "5           1  Japanese    1.0  0.742574\n",
      "6           2  Japanese    1.0  0.829630\n",
      "7           3  Japanese    1.0  0.803922\n",
      "8           4  Japanese    1.0  0.672131\n",
      "9           5  Japanese    1.0  0.803738\n",
      "10          1    Korean    1.0  0.801980\n",
      "11          2    Korean    1.0  0.725926\n",
      "12          3    Korean    1.0  0.882353\n",
      "13          4    Korean    1.0  0.573770\n",
      "14          5    Korean    1.0  0.616822\n",
      "15          1   Spanish    1.0  0.732673\n",
      "16          2   Spanish    1.0  0.414815\n",
      "17          3   Spanish    1.0  0.450980\n",
      "18          4   Spanish    1.0  0.278689\n",
      "19          5   Spanish    1.0  0.429907\n",
      "20          1       UKS    1.0  0.267327\n",
      "21          2       UKS    1.0  0.148148\n",
      "22          3       UKS    1.0  0.254902\n",
      "23          4       UKS    1.0  0.147541\n",
      "24          5       UKS    1.0  0.074766\n",
      "25          1       USS    1.0  0.306931\n",
      "26          2       USS    1.0  0.148148\n",
      "27          3       USS    1.0  0.254902\n",
      "28          4       USS    1.0  0.065574\n",
      "29          5       USS    1.0  0.074766\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-100-b1b89866beed>:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  df_accent['wer'] = accent_wer\n"
     ]
    }
   ],
   "source": [
    "df_accent['wer'] = accent_wer\n",
    "print(df_accent)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Visualize the WER for different accent (colors are for accents)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "color1 = [\"red\"]*len_origin+[\"yellow\"]*len_origin+[\"blue\"]*len_origin+[\"purple\"]*len_origin+[\"orange\"]*len_origin+[\"black\"]*len_origin"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(df_accent['Accent'], df_accent['wer'], c=color1)\n",
    "\n",
    "plt.xlabel('Accent')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.title('Accuracy of Amazon Transcribe for different accent')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<font size=\"3\"> WER for different speed </font>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Read the hypothesis text files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('speed text.txt', 'r') as reader:\n",
    "    test = reader.readlines()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Choose last 25 entries of df that contains data for different speed of speech, store it in df_speed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_speed = df[30:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Calculate WER, store the result into list speed_wer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "origin = origin_ls*int(len(df_speed)/len_origin)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "speed_wer = []\n",
    "for i in range(len(df_speed)):\n",
    "    speed_wer.append(wer(origin[i], test[i]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Add WER values stored in speed_wer to df_speed as column \"wer\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-107-7980ba22d683>:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  df_speed['wer'] = speed_wer\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Text Type</th>\n",
       "      <th>Accent</th>\n",
       "      <th>Speed</th>\n",
       "      <th>wer</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>1</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.346535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>2</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.170370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>3</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.235294</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>4</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.065574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>5</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.074766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>1</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.306931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>2</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.177778</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>3</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.254902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>4</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.065574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>5</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.084112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>1</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.306931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>2</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.148148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>3</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.254902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>4</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.065574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>5</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.074766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>1</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.306931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>2</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.162963</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>3</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.254902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>4</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.065574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>5</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.084112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>1</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.297030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>2</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.155556</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>3</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.254902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>4</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.065574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>5</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.084112</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Text Type Accent  Speed       wer\n",
       "30          1    USS    0.5  0.346535\n",
       "31          2    USS    0.5  0.170370\n",
       "32          3    USS    0.5  0.235294\n",
       "33          4    USS    0.5  0.065574\n",
       "34          5    USS    0.5  0.074766\n",
       "35          1    USS    0.8  0.306931\n",
       "36          2    USS    0.8  0.177778\n",
       "37          3    USS    0.8  0.254902\n",
       "38          4    USS    0.8  0.065574\n",
       "39          5    USS    0.8  0.084112\n",
       "40          1    USS    1.0  0.306931\n",
       "41          2    USS    1.0  0.148148\n",
       "42          3    USS    1.0  0.254902\n",
       "43          4    USS    1.0  0.065574\n",
       "44          5    USS    1.0  0.074766\n",
       "45          1    USS    1.2  0.306931\n",
       "46          2    USS    1.2  0.162963\n",
       "47          3    USS    1.2  0.254902\n",
       "48          4    USS    1.2  0.065574\n",
       "49          5    USS    1.2  0.084112\n",
       "50          1    USS    1.5  0.297030\n",
       "51          2    USS    1.5  0.155556\n",
       "52          3    USS    1.5  0.254902\n",
       "53          4    USS    1.5  0.065574\n",
       "54          5    USS    1.5  0.084112"
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_speed['wer'] = speed_wer\n",
    "df_speed"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Visualize WER for different speed of speach"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "color1 = [\"red\"]*len_origin+[\"yellow\"]*len_origin+[\"blue\"]*len_origin+[\"purple\"]*len_origin+[\"orange\"]*len_origin"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(df_speed['Speed'], df_speed['wer'], c=color1)\n",
    "\n",
    "plt.xlabel('Speed of Speech')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.title('Accuracy of Amazon Transcribe for different speed of speech')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Combine df_accent and df_speed to get the full version of WER data my_wer_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "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>Text Type</th>\n",
       "      <th>Accent</th>\n",
       "      <th>Speed</th>\n",
       "      <th>wer</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>French</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.633663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>French</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.481481</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>French</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>French</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.426230</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>French</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.457944</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.742574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.829630</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>3</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.803922</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>4</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.672131</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>5</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.803738</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>1</td>\n",
       "      <td>Korean</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.801980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>2</td>\n",
       "      <td>Korean</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.725926</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>3</td>\n",
       "      <td>Korean</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.882353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>4</td>\n",
       "      <td>Korean</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.573770</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>5</td>\n",
       "      <td>Korean</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.616822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>1</td>\n",
       "      <td>Spanish</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.732673</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>2</td>\n",
       "      <td>Spanish</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.414815</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>3</td>\n",
       "      <td>Spanish</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.450980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>4</td>\n",
       "      <td>Spanish</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.278689</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>5</td>\n",
       "      <td>Spanish</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.429907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>1</td>\n",
       "      <td>UKS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.267327</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>2</td>\n",
       "      <td>UKS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.148148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>3</td>\n",
       "      <td>UKS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.254902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>4</td>\n",
       "      <td>UKS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.147541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>5</td>\n",
       "      <td>UKS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.074766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>1</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.306931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>2</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.148148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>3</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.254902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>4</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.065574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>5</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.074766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>1</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.346535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>2</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.170370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>3</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.235294</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>4</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.065574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>5</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.074766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>1</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.306931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>2</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.177778</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>3</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.254902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>4</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.065574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>5</td>\n",
       "      <td>USS</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.084112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>1</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.306931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>2</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.148148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>3</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.254902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>4</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.065574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>5</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.074766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>1</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.306931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>2</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.162963</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>3</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.254902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>4</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.065574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>5</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.2</td>\n",
       "      <td>0.084112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>1</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.297030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>2</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.155556</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>3</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.254902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>4</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.065574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>5</td>\n",
       "      <td>USS</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.084112</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Text Type    Accent  Speed       wer\n",
       "0           1    French    1.0  0.633663\n",
       "1           2    French    1.0  0.481481\n",
       "2           3    French    1.0  0.666667\n",
       "3           4    French    1.0  0.426230\n",
       "4           5    French    1.0  0.457944\n",
       "5           1  Japanese    1.0  0.742574\n",
       "6           2  Japanese    1.0  0.829630\n",
       "7           3  Japanese    1.0  0.803922\n",
       "8           4  Japanese    1.0  0.672131\n",
       "9           5  Japanese    1.0  0.803738\n",
       "10          1    Korean    1.0  0.801980\n",
       "11          2    Korean    1.0  0.725926\n",
       "12          3    Korean    1.0  0.882353\n",
       "13          4    Korean    1.0  0.573770\n",
       "14          5    Korean    1.0  0.616822\n",
       "15          1   Spanish    1.0  0.732673\n",
       "16          2   Spanish    1.0  0.414815\n",
       "17          3   Spanish    1.0  0.450980\n",
       "18          4   Spanish    1.0  0.278689\n",
       "19          5   Spanish    1.0  0.429907\n",
       "20          1       UKS    1.0  0.267327\n",
       "21          2       UKS    1.0  0.148148\n",
       "22          3       UKS    1.0  0.254902\n",
       "23          4       UKS    1.0  0.147541\n",
       "24          5       UKS    1.0  0.074766\n",
       "25          1       USS    1.0  0.306931\n",
       "26          2       USS    1.0  0.148148\n",
       "27          3       USS    1.0  0.254902\n",
       "28          4       USS    1.0  0.065574\n",
       "29          5       USS    1.0  0.074766\n",
       "30          1       USS    0.5  0.346535\n",
       "31          2       USS    0.5  0.170370\n",
       "32          3       USS    0.5  0.235294\n",
       "33          4       USS    0.5  0.065574\n",
       "34          5       USS    0.5  0.074766\n",
       "35          1       USS    0.8  0.306931\n",
       "36          2       USS    0.8  0.177778\n",
       "37          3       USS    0.8  0.254902\n",
       "38          4       USS    0.8  0.065574\n",
       "39          5       USS    0.8  0.084112\n",
       "40          1       USS    1.0  0.306931\n",
       "41          2       USS    1.0  0.148148\n",
       "42          3       USS    1.0  0.254902\n",
       "43          4       USS    1.0  0.065574\n",
       "44          5       USS    1.0  0.074766\n",
       "45          1       USS    1.2  0.306931\n",
       "46          2       USS    1.2  0.162963\n",
       "47          3       USS    1.2  0.254902\n",
       "48          4       USS    1.2  0.065574\n",
       "49          5       USS    1.2  0.084112\n",
       "50          1       USS    1.5  0.297030\n",
       "51          2       USS    1.5  0.155556\n",
       "52          3       USS    1.5  0.254902\n",
       "53          4       USS    1.5  0.065574\n",
       "54          5       USS    1.5  0.084112"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my_wer_df = df_accent.append(df_speed)\n",
    "my_wer_df"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}