3_analyze_annotated.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Analyze the results from annotations\n",
    "Show agreements and disagreements between the annotated polarities and the document- sentence- and target- level annotations\n",
    "\n",
    "Note: This repositiry contains code that is being updated as the project is proceeding, and adjustments may have occured since the publishing of the ELSA paper"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import os\n",
    "import time\n",
    "from collections import Counter, defaultdict\n",
    "import numpy as np\n",
    "import random\n",
    "import pandas as pd\n",
    "import torch\n",
    "\n",
    "save_root = \"outputs/\"\n",
    "tabular_savefolder = os.path.join(save_root, \"tabular\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Document polarity vs entity polarity: 133 281 0.473\n",
      "doc_pol     Negative  Neutral  Positive\n",
      "manual_pol                             \n",
      "Neg               15        8         5\n",
      "Neu               29       66        75\n",
      "Pos                3       26        52\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='doc_pol'>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "\n",
    "excel_path = os.path.join(tabular_savefolder,\"elsa_entities_annotated.xlsx\" )\n",
    "df = pd.read_excel(excel_path)\n",
    "df = df.loc[df[\"manual_pol\"]!=\"Spurious\"]\n",
    "\n",
    "doc_map = {1:\"Negative\", 2:\"Negative\", 3:\"Neutral\", 4:\"Neutral\", 5:\"Positive\", 6:\"Positive\"}\n",
    "df[\"doc_pol\"] = df[\"doc_rating\"].map(doc_map)\n",
    "doc_elsa = pd.pivot_table(df, index = [\"manual_pol\"], values = \"entity_id\", columns = \"doc_pol\", aggfunc='count')\n",
    "# Document-level polarity vs entity polarity\n",
    "hits = np.sum([doc_elsa.at[pol[:3], pol] for pol in [\"Negative\", \"Neutral\",\"Positive\"]])\n",
    "print(\"Document polarity vs entity polarity:\",hits, len(df), round(hits/len(df), 3))\n",
    "\n",
    "print(doc_elsa)\n",
    "doc_elsa.T.plot.bar(stacked=True)\n",
    "for c in doc_elsa.columns:\n",
    "    doc_elsa[c+\"_rel\"] = doc_elsa[c]/doc_elsa[c].sum()\n",
    "doc_elsa[[c for c in doc_elsa.columns if \"_rel\" in c]].T.plot.bar(stacked=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sentence polarity vs entity polarity: 194.0 281 0.69\n",
      "sent_pol_resolved  Mixed  Negative  Neutral  Positive\n",
      "manual_pol                                           \n",
      "Neg                  9.0      10.0      8.0       1.0\n",
      "Neu                  4.0      17.0    115.0      34.0\n",
      "Pos                  9.0       NaN      3.0      69.0\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='sent_pol_resolved'>"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Sentence polarity\n",
    "sent_elsa = pd.pivot_table(df, index = [\"manual_pol\"], values = \"entity_id\", columns = \"sent_pol_resolved\", aggfunc='count')\n",
    "# Sentence-level polarity vs entity polarity\n",
    "hits = np.sum([sent_elsa.at[pol[:3], pol] for pol in [\"Negative\", \"Neutral\",\"Positive\"]])\n",
    "print(\"Sentence polarity vs entity polarity:\",hits, len(df), round(hits/len(df), 3))\n",
    "\n",
    "print(sent_elsa)\n",
    "sent_elsa.T.plot.bar(stacked=True)\n",
    "for c in sent_elsa.columns:\n",
    "    sent_elsa[c+\"_rel\"] = sent_elsa[c]/sent_elsa[c].sum()\n",
    "sent_elsa[[c for c in sent_elsa.columns if \"_rel\" in c]].T.plot.bar(stacked=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sentiment target polarity vs entity polarity: 231.0 281 0.822\n",
      "targ_pol_resolved  Mixed  Negative  Neutral  Positive\n",
      "manual_pol                                           \n",
      "Neg                  1.0      13.0     10.0       4.0\n",
      "Neu                  0.0       0.0    164.0       6.0\n",
      "Pos                  4.0       3.0     20.0      54.0\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='targ_pol_resolved'>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# ELSA vs annotated sentiment targets\n",
    "df[\"targ_pol_resolved\"]= df[\"targ_pol_resolved\"].fillna(\"Neutral\")\n",
    "targets_elsa = pd.pivot_table(df, index = [\"manual_pol\"], values = \"entity_id\", columns = \"targ_pol_resolved\", aggfunc='count').fillna(0)\n",
    "# Sentence-level polarity vs entity polarity\n",
    "hits = np.sum([targets_elsa.at[pol[:3], pol] for pol in [\"Negative\", \"Neutral\",\"Positive\"]])\n",
    "print(\"Sentiment target polarity vs entity polarity:\",hits, len(df), round(hits/len(df), 3))\n",
    "\n",
    "print(targets_elsa)\n",
    "targets_elsa.T.plot.bar(stacked=True)\n",
    "for c in targets_elsa.columns:\n",
    "    targets_elsa[c+\"_rel\"] = targets_elsa[c]/targets_elsa[c].sum()\n",
    "targets_elsa[[c for c in targets_elsa.columns if \"_rel\" in c]].T.plot.bar(stacked=True)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9.7 ('phdev')",
   "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": "3.9.7"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "700a218f876b011520bcab0bf4c4fd60938a4cf5c91c77ab574cb3838b05995c"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}