seaboard_plot_results.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [],
   "source": [
    "header = ['dataset', 'lr', 'num_epochs', 'oversampling_ratio', \\\n",
    " 'num_layers', 'dropout', 'model_type', 'batch_size', 'only_binary', \\\n",
    "'exp_name', 'standardize', 'features', 'time_cutoff', 'seed', 'hidden_dim']\n",
    "header_acc = [\"train_acc\", \"val_acc\", \"test_acc\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 257,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"new_results.csv\")\n",
    "df_temp = df.copy()\n",
    "df_temp[\"index\"] = df_temp.index\n",
    "df_temp_melt = pd.melt(df_temp, value_name=\"Accuracy\", value_vars=header_acc, id_vars=header+[\"index\"], var_name=\"split\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a2da846d8>"
      ]
     },
     "execution_count": 258,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "sns.scatterplot(data=df_temp_melt, x=\"index\", y=\"Accuracy\", hue=\"split\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_final = df.drop_duplicates(subset=header, keep=\"last\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dataset ['twitter15' 'twitter16']\n",
      "lr [0.01]\n",
      "num_epochs [500]\n",
      "oversampling_ratio [1]\n",
      "num_layers [2]\n",
      "dropout [0.]\n",
      "model_type ['GAT' 'GraphSage' 'GCN']\n",
      "batch_size [32]\n",
      "only_binary [False]\n",
      "exp_name ['final' 'final_text_only_GAT' 'final_user_only_GAT'\n",
      " 'final15_all_GraphSage' 'final15_text_only_GraphSage'\n",
      " 'final15_user_only_GraphSage' 'final15_all_GCN' 'final16_all_GCN'\n",
      " 'final15_text_only_GCN' 'final16_text_only_GCN' 'final15_user_only_GCN'\n",
      " 'final16_user_only_GCN' 'final16_all_GAT' 'final16_text_only_GAT'\n",
      " 'final16_user_only_GAT' 'final16_all_GraphSage'\n",
      " 'final16_text_only_GraphSage' 'final16_user_only_GraphSage']\n",
      "standardize [ True]\n",
      "features ['all' 'text_only' 'user_only']\n",
      "time_cutoff ['None']\n",
      "seed [64 43]\n",
      "hidden_dim [64]\n",
      "train_acc [0.99900498 0.99701493 0.87761194 0.93930348 0.96218905 0.67860697\n",
      " 0.85869565 0.99800995 0.92028986 0.71442786 0.88586957 0.98369565\n",
      " 0.90036232 0.90942029 0.96376812 0.86594203 0.80253623]\n",
      "val_acc [0.67114094 0.69127517 0.65771812 0.67785235 0.63758389 0.71812081\n",
      " 0.84146341 0.66442953 0.73170732 0.80487805 0.74390244 0.68292683\n",
      " 0.75609756 0.92682927]\n",
      "test_acc [0.66071429 0.63095238 0.60119048 0.65178571 0.5922619  0.54761905\n",
      " 0.69047619 0.75       0.55978261 0.61904762 0.625      0.68478261\n",
      " 0.57065217 0.66847826 0.86413043 0.71195652]\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>lr</th>\n",
       "      <th>num_epochs</th>\n",
       "      <th>oversampling_ratio</th>\n",
       "      <th>num_layers</th>\n",
       "      <th>dropout</th>\n",
       "      <th>batch_size</th>\n",
       "      <th>seed</th>\n",
       "      <th>hidden_dim</th>\n",
       "      <th>train_acc</th>\n",
       "      <th>val_acc</th>\n",
       "      <th>test_acc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1.800000e+01</td>\n",
       "      <td>18.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.0</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.000000e-02</td>\n",
       "      <td>500.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>57.000000</td>\n",
       "      <td>64.0</td>\n",
       "      <td>0.903097</td>\n",
       "      <td>0.723477</td>\n",
       "      <td>0.647293</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1.785016e-18</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10.186496</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.094543</td>\n",
       "      <td>0.075035</td>\n",
       "      <td>0.077363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.000000e-02</td>\n",
       "      <td>500.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>43.000000</td>\n",
       "      <td>64.0</td>\n",
       "      <td>0.678607</td>\n",
       "      <td>0.637584</td>\n",
       "      <td>0.547619</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.000000e-02</td>\n",
       "      <td>500.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>43.000000</td>\n",
       "      <td>64.0</td>\n",
       "      <td>0.868860</td>\n",
       "      <td>0.672819</td>\n",
       "      <td>0.594494</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.000000e-02</td>\n",
       "      <td>500.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>64.0</td>\n",
       "      <td>0.914855</td>\n",
       "      <td>0.704698</td>\n",
       "      <td>0.641369</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.000000e-02</td>\n",
       "      <td>500.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>64.0</td>\n",
       "      <td>0.978714</td>\n",
       "      <td>0.753049</td>\n",
       "      <td>0.680707</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.000000e-02</td>\n",
       "      <td>500.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>64.0</td>\n",
       "      <td>0.999005</td>\n",
       "      <td>0.926829</td>\n",
       "      <td>0.864130</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 lr  num_epochs  oversampling_ratio  num_layers  dropout  \\\n",
       "count  1.800000e+01        18.0                18.0        18.0     18.0   \n",
       "mean   1.000000e-02       500.0                 1.0         2.0      0.0   \n",
       "std    1.785016e-18         0.0                 0.0         0.0      0.0   \n",
       "min    1.000000e-02       500.0                 1.0         2.0      0.0   \n",
       "25%    1.000000e-02       500.0                 1.0         2.0      0.0   \n",
       "50%    1.000000e-02       500.0                 1.0         2.0      0.0   \n",
       "75%    1.000000e-02       500.0                 1.0         2.0      0.0   \n",
       "max    1.000000e-02       500.0                 1.0         2.0      0.0   \n",
       "\n",
       "       batch_size       seed  hidden_dim  train_acc    val_acc   test_acc  \n",
       "count        18.0  18.000000        18.0  18.000000  18.000000  18.000000  \n",
       "mean         32.0  57.000000        64.0   0.903097   0.723477   0.647293  \n",
       "std           0.0  10.186496         0.0   0.094543   0.075035   0.077363  \n",
       "min          32.0  43.000000        64.0   0.678607   0.637584   0.547619  \n",
       "25%          32.0  43.000000        64.0   0.868860   0.672819   0.594494  \n",
       "50%          32.0  64.000000        64.0   0.914855   0.704698   0.641369  \n",
       "75%          32.0  64.000000        64.0   0.978714   0.753049   0.680707  \n",
       "max          32.0  64.000000        64.0   0.999005   0.926829   0.864130  "
      ]
     },
     "execution_count": 260,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for col in df_final.columns:\n",
    "    print(col, df_final[col].unique())\n",
    "df_final.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 261,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_melt = pd.melt(df_final, value_name=\"Accuracy\", value_vars=header_acc, id_vars=header, var_name=\"split\")\n",
    "df_melt[\"split\"] = [x.split(\"_\")[0] for x in df_melt[\"split\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 262,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_melt[\"Model\"] = df_melt[\"model_type\"] + \"\\n\" + df_melt[\"features\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 263,
   "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>dataset</th>\n",
       "      <th>lr</th>\n",
       "      <th>num_epochs</th>\n",
       "      <th>oversampling_ratio</th>\n",
       "      <th>num_layers</th>\n",
       "      <th>dropout</th>\n",
       "      <th>model_type</th>\n",
       "      <th>batch_size</th>\n",
       "      <th>only_binary</th>\n",
       "      <th>exp_name</th>\n",
       "      <th>standardize</th>\n",
       "      <th>features</th>\n",
       "      <th>time_cutoff</th>\n",
       "      <th>seed</th>\n",
       "      <th>hidden_dim</th>\n",
       "      <th>split</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Model</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>twitter15</td>\n",
       "      <td>0.01</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GCN</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>final15_all_GCN</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64</td>\n",
       "      <td>64</td>\n",
       "      <td>train</td>\n",
       "      <td>0.999005</td>\n",
       "      <td>GCN\\nall</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.01</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GCN</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>final16_all_GCN</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>43</td>\n",
       "      <td>64</td>\n",
       "      <td>train</td>\n",
       "      <td>0.858696</td>\n",
       "      <td>GCN\\nall</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>twitter15</td>\n",
       "      <td>0.01</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GCN</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>final15_all_GCN</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64</td>\n",
       "      <td>64</td>\n",
       "      <td>val</td>\n",
       "      <td>0.718121</td>\n",
       "      <td>GCN\\nall</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.01</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GCN</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>final16_all_GCN</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>43</td>\n",
       "      <td>64</td>\n",
       "      <td>val</td>\n",
       "      <td>0.841463</td>\n",
       "      <td>GCN\\nall</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>twitter15</td>\n",
       "      <td>0.01</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GCN</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>final15_all_GCN</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64</td>\n",
       "      <td>64</td>\n",
       "      <td>test</td>\n",
       "      <td>0.690476</td>\n",
       "      <td>GCN\\nall</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.01</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GCN</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>final16_all_GCN</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>43</td>\n",
       "      <td>64</td>\n",
       "      <td>test</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>GCN\\nall</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      dataset    lr  num_epochs  oversampling_ratio  num_layers  dropout  \\\n",
       "6   twitter15  0.01         500                   1           2      0.0   \n",
       "7   twitter16  0.01         500                   1           2      0.0   \n",
       "24  twitter15  0.01         500                   1           2      0.0   \n",
       "25  twitter16  0.01         500                   1           2      0.0   \n",
       "42  twitter15  0.01         500                   1           2      0.0   \n",
       "43  twitter16  0.01         500                   1           2      0.0   \n",
       "\n",
       "   model_type  batch_size  only_binary         exp_name  standardize features  \\\n",
       "6         GCN          32        False  final15_all_GCN         True      all   \n",
       "7         GCN          32        False  final16_all_GCN         True      all   \n",
       "24        GCN          32        False  final15_all_GCN         True      all   \n",
       "25        GCN          32        False  final16_all_GCN         True      all   \n",
       "42        GCN          32        False  final15_all_GCN         True      all   \n",
       "43        GCN          32        False  final16_all_GCN         True      all   \n",
       "\n",
       "   time_cutoff  seed  hidden_dim  split  Accuracy     Model  \n",
       "6         None    64          64  train  0.999005  GCN\\nall  \n",
       "7         None    43          64  train  0.858696  GCN\\nall  \n",
       "24        None    64          64    val  0.718121  GCN\\nall  \n",
       "25        None    43          64    val  0.841463  GCN\\nall  \n",
       "42        None    64          64   test  0.690476  GCN\\nall  \n",
       "43        None    43          64   test  0.750000  GCN\\nall  "
      ]
     },
     "execution_count": 263,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_melt[df_melt.Model==\"GCN\" + \"\\n\" + \"all\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(15, 4))\n",
    "sns.barplot(data=df_melt[df_melt.dataset==\"twitter15\"], y=\"Accuracy\", x=\"Model\", hue=\"split\")\n",
    "plt.ylim([0.5, None]);\n",
    "plt.title(\"Results on Twitter15\")\n",
    "plt.grid()\n",
    "plt.tight_layout()\n",
    "plt.savefig(\"grid15_feature_model.png\", transparent=True, dpi=400)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(15, 4))\n",
    "sns.barplot(data=df_melt[df_melt.dataset==\"twitter16\"], y=\"Accuracy\", x=\"Model\", hue=\"split\")\n",
    "plt.ylim([0.5, None]);\n",
    "plt.title(\"Results on Twitter16\")\n",
    "plt.grid()\n",
    "plt.tight_layout()\n",
    "plt.savefig(\"grid16_feature_model.png\", transparent=True, dpi=400)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "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": [
    "sns.barplot(data=df_melt[df_melt.dataset==\"twitter16\"], y=\"Accuracy\", x=\"Model\", hue=\"split\")\n",
    "plt.ylim([0.5, None]);\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Could not interpret input 'index'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-153-e7a78ffcad74>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0msns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mscatterplot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdf_melt\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"index\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"Accuracy\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhue\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"split\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/seaborn/relational.py\u001b[0m in \u001b[0;36mscatterplot\u001b[0;34m(x, y, hue, style, size, data, palette, hue_order, hue_norm, sizes, size_order, size_norm, markers, style_order, x_bins, y_bins, units, estimator, ci, n_boot, alpha, x_jitter, y_jitter, legend, ax, **kwargs)\u001b[0m\n\u001b[1;32m   1333\u001b[0m         \u001b[0mx_bins\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mx_bins\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_bins\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0my_bins\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1334\u001b[0m         \u001b[0mestimator\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mestimator\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mci\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mci\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mn_boot\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mn_boot\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1335\u001b[0;31m         \u001b[0malpha\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0malpha\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx_jitter\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mx_jitter\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_jitter\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0my_jitter\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlegend\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlegend\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1336\u001b[0m     )\n\u001b[1;32m   1337\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/seaborn/relational.py\u001b[0m in \u001b[0;36m__init__\u001b[0;34m(self, x, y, hue, size, style, data, palette, hue_order, hue_norm, sizes, size_order, size_norm, dashes, markers, style_order, x_bins, y_bins, units, estimator, ci, n_boot, alpha, x_jitter, y_jitter, legend)\u001b[0m\n\u001b[1;32m    850\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    851\u001b[0m         plot_data = self.establish_variables(\n\u001b[0;32m--> 852\u001b[0;31m             \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msize\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstyle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0munits\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    853\u001b[0m         )\n\u001b[1;32m    854\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/seaborn/relational.py\u001b[0m in \u001b[0;36mestablish_variables\u001b[0;34m(self, x, y, hue, size, style, units, data)\u001b[0m\n\u001b[1;32m    140\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstring_types\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    141\u001b[0m                     \u001b[0merr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m\"Could not interpret input '{}'\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 142\u001b[0;31m                     \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0merr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    143\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    144\u001b[0m             \u001b[0;31m# Extract variable names\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: Could not interpret input 'index'"
     ]
    }
   ],
   "source": [
    "sns.scatterplot(data=df_melt, x=\"index\", y=\"Accuracy\", hue=\"split\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "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>dataset</th>\n",
       "      <th>lr</th>\n",
       "      <th>num_epochs</th>\n",
       "      <th>oversampling_ratio</th>\n",
       "      <th>num_layers</th>\n",
       "      <th>dropout</th>\n",
       "      <th>model_type</th>\n",
       "      <th>batch_size</th>\n",
       "      <th>only_binary</th>\n",
       "      <th>exp_name</th>\n",
       "      <th>standardize</th>\n",
       "      <th>features</th>\n",
       "      <th>time_cutoff</th>\n",
       "      <th>hidden_dim</th>\n",
       "      <th>train_acc</th>\n",
       "      <th>val_acc</th>\n",
       "      <th>test_acc</th>\n",
       "      <th>index</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0100</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only_500_16_2</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.219512</td>\n",
       "      <td>0.244565</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0100</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only_500_16_2</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only</td>\n",
       "      <td>0.5778985507246377</td>\n",
       "      <td>0.621951</td>\n",
       "      <td>0.516304</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0100</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only_500_16_2</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only</td>\n",
       "      <td>0.8804347826086957</td>\n",
       "      <td>0.646341</td>\n",
       "      <td>0.635870</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0100</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only_500_16_2</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only</td>\n",
       "      <td>0.8442028985507246</td>\n",
       "      <td>0.658537</td>\n",
       "      <td>0.603261</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0100</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only_500_16_2</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only</td>\n",
       "      <td>0.8985507246376812</td>\n",
       "      <td>0.682927</td>\n",
       "      <td>0.625000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0100</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only_500_16_2</td>\n",
       "      <td>False</td>\n",
       "      <td>text_only</td>\n",
       "      <td>0.967391304347826</td>\n",
       "      <td>0.841463</td>\n",
       "      <td>0.673913</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.393116</td>\n",
       "      <td>0.304878</td>\n",
       "      <td>0.347826</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.467391</td>\n",
       "      <td>0.365854</td>\n",
       "      <td>0.391304</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.523551</td>\n",
       "      <td>0.390244</td>\n",
       "      <td>0.434783</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.532609</td>\n",
       "      <td>0.402439</td>\n",
       "      <td>0.467391</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.543478</td>\n",
       "      <td>0.439024</td>\n",
       "      <td>0.456522</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.559783</td>\n",
       "      <td>0.451220</td>\n",
       "      <td>0.472826</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.574275</td>\n",
       "      <td>0.463415</td>\n",
       "      <td>0.483696</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.572464</td>\n",
       "      <td>0.475610</td>\n",
       "      <td>0.472826</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.561594</td>\n",
       "      <td>0.487805</td>\n",
       "      <td>0.478261</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.668478</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.532609</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.713768</td>\n",
       "      <td>0.524390</td>\n",
       "      <td>0.538043</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.719203</td>\n",
       "      <td>0.548780</td>\n",
       "      <td>0.527174</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.742754</td>\n",
       "      <td>0.560976</td>\n",
       "      <td>0.565217</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0.573171</td>\n",
       "      <td>0.614130</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.760870</td>\n",
       "      <td>0.585366</td>\n",
       "      <td>0.576087</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.766304</td>\n",
       "      <td>0.597561</td>\n",
       "      <td>0.554348</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.793478</td>\n",
       "      <td>0.609756</td>\n",
       "      <td>0.608696</td>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.807971</td>\n",
       "      <td>0.634146</td>\n",
       "      <td>0.586957</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.840580</td>\n",
       "      <td>0.658537</td>\n",
       "      <td>0.597826</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.862319</td>\n",
       "      <td>0.670732</td>\n",
       "      <td>0.597826</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.885870</td>\n",
       "      <td>0.707317</td>\n",
       "      <td>0.635870</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.938406</td>\n",
       "      <td>0.731707</td>\n",
       "      <td>0.641304</td>\n",
       "      <td>27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.942029</td>\n",
       "      <td>0.743902</td>\n",
       "      <td>0.690217</td>\n",
       "      <td>28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.756098</td>\n",
       "      <td>0.695652</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.967391</td>\n",
       "      <td>0.768293</td>\n",
       "      <td>0.663043</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>twitter16</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>500</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.1</td>\n",
       "      <td>GAT</td>\n",
       "      <td>32</td>\n",
       "      <td>False</td>\n",
       "      <td>all_500_16_2_test</td>\n",
       "      <td>True</td>\n",
       "      <td>all</td>\n",
       "      <td>None</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>0.994565</td>\n",
       "      <td>0.792683</td>\n",
       "      <td>0.679348</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      dataset      lr  num_epochs  oversampling_ratio  num_layers  dropout  \\\n",
       "0   twitter16  0.0100         500                   1           2      0.0   \n",
       "1   twitter16  0.0100         500                   1           2      0.0   \n",
       "2   twitter16  0.0100         500                   1           2      0.0   \n",
       "3   twitter16  0.0100         500                   1           2      0.0   \n",
       "4   twitter16  0.0100         500                   1           2      0.0   \n",
       "5   twitter16  0.0100         500                   1           2      0.0   \n",
       "6   twitter16  0.0001         500                   1           2      0.1   \n",
       "7   twitter16  0.0001         500                   1           2      0.1   \n",
       "8   twitter16  0.0001         500                   1           2      0.1   \n",
       "9   twitter16  0.0001         500                   1           2      0.1   \n",
       "10  twitter16  0.0001         500                   1           2      0.1   \n",
       "11  twitter16  0.0001         500                   1           2      0.1   \n",
       "12  twitter16  0.0001         500                   1           2      0.1   \n",
       "13  twitter16  0.0001         500                   1           2      0.1   \n",
       "14  twitter16  0.0001         500                   1           2      0.1   \n",
       "15  twitter16  0.0001         500                   1           2      0.1   \n",
       "16  twitter16  0.0001         500                   1           2      0.1   \n",
       "17  twitter16  0.0001         500                   1           2      0.1   \n",
       "18  twitter16  0.0001         500                   1           2      0.1   \n",
       "19  twitter16  0.0001         500                   1           2      0.1   \n",
       "20  twitter16  0.0001         500                   1           2      0.1   \n",
       "21  twitter16  0.0001         500                   1           2      0.1   \n",
       "22  twitter16  0.0001         500                   1           2      0.1   \n",
       "23  twitter16  0.0001         500                   1           2      0.1   \n",
       "24  twitter16  0.0001         500                   1           2      0.1   \n",
       "25  twitter16  0.0001         500                   1           2      0.1   \n",
       "26  twitter16  0.0001         500                   1           2      0.1   \n",
       "27  twitter16  0.0001         500                   1           2      0.1   \n",
       "28  twitter16  0.0001         500                   1           2      0.1   \n",
       "29  twitter16  0.0001         500                   1           2      0.1   \n",
       "30  twitter16  0.0001         500                   1           2      0.1   \n",
       "31  twitter16  0.0001         500                   1           2      0.1   \n",
       "\n",
       "   model_type  batch_size  only_binary            exp_name  standardize  \\\n",
       "0         GAT          32        False  text_only_500_16_2        False   \n",
       "1         GAT          32        False  text_only_500_16_2        False   \n",
       "2         GAT          32        False  text_only_500_16_2        False   \n",
       "3         GAT          32        False  text_only_500_16_2        False   \n",
       "4         GAT          32        False  text_only_500_16_2        False   \n",
       "5         GAT          32        False  text_only_500_16_2        False   \n",
       "6         GAT          32        False   all_500_16_2_test         True   \n",
       "7         GAT          32        False   all_500_16_2_test         True   \n",
       "8         GAT          32        False   all_500_16_2_test         True   \n",
       "9         GAT          32        False   all_500_16_2_test         True   \n",
       "10        GAT          32        False   all_500_16_2_test         True   \n",
       "11        GAT          32        False   all_500_16_2_test         True   \n",
       "12        GAT          32        False   all_500_16_2_test         True   \n",
       "13        GAT          32        False   all_500_16_2_test         True   \n",
       "14        GAT          32        False   all_500_16_2_test         True   \n",
       "15        GAT          32        False   all_500_16_2_test         True   \n",
       "16        GAT          32        False   all_500_16_2_test         True   \n",
       "17        GAT          32        False   all_500_16_2_test         True   \n",
       "18        GAT          32        False   all_500_16_2_test         True   \n",
       "19        GAT          32        False   all_500_16_2_test         True   \n",
       "20        GAT          32        False   all_500_16_2_test         True   \n",
       "21        GAT          32        False   all_500_16_2_test         True   \n",
       "22        GAT          32        False   all_500_16_2_test         True   \n",
       "23        GAT          32        False   all_500_16_2_test         True   \n",
       "24        GAT          32        False   all_500_16_2_test         True   \n",
       "25        GAT          32        False   all_500_16_2_test         True   \n",
       "26        GAT          32        False   all_500_16_2_test         True   \n",
       "27        GAT          32        False   all_500_16_2_test         True   \n",
       "28        GAT          32        False   all_500_16_2_test         True   \n",
       "29        GAT          32        False   all_500_16_2_test         True   \n",
       "30        GAT          32        False   all_500_16_2_test         True   \n",
       "31        GAT          32        False   all_500_16_2_test         True   \n",
       "\n",
       "     features         time_cutoff  hidden_dim  train_acc   val_acc  test_acc  \\\n",
       "0   text_only                0.25    0.219512   0.244565       NaN       NaN   \n",
       "1   text_only  0.5778985507246377    0.621951   0.516304       NaN       NaN   \n",
       "2   text_only  0.8804347826086957    0.646341   0.635870       NaN       NaN   \n",
       "3   text_only  0.8442028985507246    0.658537   0.603261       NaN       NaN   \n",
       "4   text_only  0.8985507246376812    0.682927   0.625000       NaN       NaN   \n",
       "5   text_only   0.967391304347826    0.841463   0.673913       NaN       NaN   \n",
       "6         all                None   64.000000   0.393116  0.304878  0.347826   \n",
       "7         all                None   64.000000   0.467391  0.365854  0.391304   \n",
       "8         all                None   64.000000   0.523551  0.390244  0.434783   \n",
       "9         all                None   64.000000   0.532609  0.402439  0.467391   \n",
       "10        all                None   64.000000   0.543478  0.439024  0.456522   \n",
       "11        all                None   64.000000   0.559783  0.451220  0.472826   \n",
       "12        all                None   64.000000   0.574275  0.463415  0.483696   \n",
       "13        all                None   64.000000   0.572464  0.475610  0.472826   \n",
       "14        all                None   64.000000   0.561594  0.487805  0.478261   \n",
       "15        all                None   64.000000   0.668478  0.500000  0.532609   \n",
       "16        all                None   64.000000   0.713768  0.524390  0.538043   \n",
       "17        all                None   64.000000   0.719203  0.548780  0.527174   \n",
       "18        all                None   64.000000   0.742754  0.560976  0.565217   \n",
       "19        all                None   64.000000   0.750000  0.573171  0.614130   \n",
       "20        all                None   64.000000   0.760870  0.585366  0.576087   \n",
       "21        all                None   64.000000   0.766304  0.597561  0.554348   \n",
       "22        all                None   64.000000   0.793478  0.609756  0.608696   \n",
       "23        all                None   64.000000   0.807971  0.634146  0.586957   \n",
       "24        all                None   64.000000   0.840580  0.658537  0.597826   \n",
       "25        all                None   64.000000   0.862319  0.670732  0.597826   \n",
       "26        all                None   64.000000   0.885870  0.707317  0.635870   \n",
       "27        all                None   64.000000   0.938406  0.731707  0.641304   \n",
       "28        all                None   64.000000   0.942029  0.743902  0.690217   \n",
       "29        all                None   64.000000   0.958333  0.756098  0.695652   \n",
       "30        all                None   64.000000   0.967391  0.768293  0.663043   \n",
       "31        all                None   64.000000   0.994565  0.792683  0.679348   \n",
       "\n",
       "    index  \n",
       "0       0  \n",
       "1       1  \n",
       "2       2  \n",
       "3       3  \n",
       "4       4  \n",
       "5       5  \n",
       "6       6  \n",
       "7       7  \n",
       "8       8  \n",
       "9       9  \n",
       "10     10  \n",
       "11     11  \n",
       "12     12  \n",
       "13     13  \n",
       "14     14  \n",
       "15     15  \n",
       "16     16  \n",
       "17     17  \n",
       "18     18  \n",
       "19     19  \n",
       "20     20  \n",
       "21     21  \n",
       "22     22  \n",
       "23     23  \n",
       "24     24  \n",
       "25     25  \n",
       "26     26  \n",
       "27     27  \n",
       "28     28  \n",
       "29     29  \n",
       "30     30  \n",
       "31     31  "
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Record (best val acc)')"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "sns.scatterplot(data=df_melt16, x=\"index\", y=\"Accuracy\", hue=\"split\", style=\"features\")\n",
    "plt.title(\"Twitter16 Accuracy\")\n",
    "plt.xlabel(\"Record (best val acc)\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_early = pd.read_csv(\"earlydetec_results.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_early = df_early.drop_duplicates(subset=header, keep=\"last\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_melt = pd.melt(df_early, value_name=\"Accuracy\", value_vars=header_acc, id_vars=header, var_name=\"split\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_melt[\"time_cutoff\"] = df_melt[\"time_cutoff\"].astype(float)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15, 6))\n",
    "sns.lineplot(data=df_melt[df_melt.features==\"all\"], y=\"Accuracy\", x=\"time_cutoff\", hue=\"split\")\n",
    "plt.xticks(rotation=90);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}