Individual-level models for mapping approach.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import pickle\n",
    "import numpy as np\n",
    "from sklearn.model_selection import LeaveOneOut\n",
    "from sklearn.linear_model import Ridge\n",
    "from scipy.stats import pearsonr\n",
    "from tqdm import tqdm\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "domains = ['brand', 'good', 'trait', 'food', 'occupation', 'risk', 'people']\n",
    "dims = [('sincere','exciting'),\n",
    "        ('hedonic','utilitarian'),\n",
    "        ('masculine','feminine'),\n",
    "        ('tasty','nutritious'),\n",
    "        ('significance','autonomy'),\n",
    "        ('dread-inducing','unknowable'),\n",
    "        ('warm', 'competent')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"data/dict_of_Xs.pickle\", \"rb\") as handle:\n",
    "    dict_of_Xs = pickle.load(handle)\n",
    "    \n",
    "with open(\"data/dict_of_individual_ys.pickle\", \"rb\") as handle:\n",
    "    dict_of_individual_ys = pickle.load(handle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "ridge = Ridge(alpha=10) # pre-reg'd analyses revealed ridge with alpha=10 had highest rsquared across dimensions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "dict_of_individual_y_preds = dict()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Check dimension-specific distributions of percentage of items not rated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>dim</th>\n",
       "      <th>mean_missing_items</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>sincere</td>\n",
       "      <td>20.882353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>exciting</td>\n",
       "      <td>20.882353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>hedonic</td>\n",
       "      <td>4.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>utilitarian</td>\n",
       "      <td>3.282609</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>masculine</td>\n",
       "      <td>26.863636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>feminine</td>\n",
       "      <td>26.909091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>tasty</td>\n",
       "      <td>9.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>nutritious</td>\n",
       "      <td>8.437500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>significance</td>\n",
       "      <td>6.809524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>autonomy</td>\n",
       "      <td>6.809524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>dread-inducing</td>\n",
       "      <td>3.051282</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>unknowable</td>\n",
       "      <td>5.076923</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>warm</td>\n",
       "      <td>22.655172</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>competent</td>\n",
       "      <td>20.551724</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               dim  mean_missing_items\n",
       "0          sincere           20.882353\n",
       "1         exciting           20.882353\n",
       "2          hedonic            4.000000\n",
       "3      utilitarian            3.282609\n",
       "4        masculine           26.863636\n",
       "5         feminine           26.909091\n",
       "6            tasty            9.500000\n",
       "7       nutritious            8.437500\n",
       "8     significance            6.809524\n",
       "9         autonomy            6.809524\n",
       "10  dread-inducing            3.051282\n",
       "11      unknowable            5.076923\n",
       "12            warm           22.655172\n",
       "13       competent           20.551724"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mean_missing = pd.DataFrame([(dim, df.isna().sum(axis='columns').mean()) for dim, df in dict_of_individual_ys.items()], \n",
    "                             columns=['dim','mean_missing_items'])\n",
    "mean_missing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "mean_missing_items    13.265121\n",
       "dtype: float64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mean_missing.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "7it [01:50, 15.80s/it]\n"
     ]
    }
   ],
   "source": [
    "for domain, dim_pair in tqdm(zip(domains, dims)):\n",
    "    X = dict_of_Xs[domain].values\n",
    "#     X = X[:100]\n",
    "    for dim in dim_pair:\n",
    "        all_individual_preds = np.zeros(shape=dict_of_individual_ys[dim].shape)\n",
    "#         all_individual_preds = np.zeros(shape=(dict_of_individual_ys[dim].shape[0], 100))\n",
    "        for individual_idx, y in enumerate(dict_of_individual_ys[dim].values):\n",
    "#             y = y[:100]\n",
    "            y_preds = np.zeros(shape=len(y))\n",
    "            loo = LeaveOneOut()\n",
    "            for train_index, test_index in loo.split(X):\n",
    "                X_train, X_test = X[train_index], X[test_index]\n",
    "                y_train = y[train_index].astype(float)\n",
    "\n",
    "                y_notna_idxs = ~np.isnan(y_train)\n",
    "                y_train = y_train[y_notna_idxs]\n",
    "                X_train = X_train[y_notna_idxs]\n",
    "\n",
    "                ridge.fit(X=X_train, y=y_train)\n",
    "                y_pred = ridge.predict(X=X_test)[0]\n",
    "                y_preds[test_index] = y_pred\n",
    "            all_individual_preds[individual_idx] = y_preds\n",
    "        individual_predictions = pd.DataFrame(all_individual_preds, columns=dict_of_individual_ys[dim].columns)\n",
    "#         individual_predictions = pd.DataFrame(all_individual_preds, columns=dict_of_individual_ys[dim].columns[:100])\n",
    "        dict_of_individual_y_preds[dim] = individual_predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"data/dict_of_individual_y_preds.pickle\", \"wb\") as handle:\n",
    "    pickle.dump(dict_of_individual_y_preds, handle, protocol=pickle.HIGHEST_PROTOCOL)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Compute and visualize actual vs pred correlations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "if 'dict_of_individual_y_preds' not in globals():\n",
    "    with open(\"data/dict_of_individual_y_preds.pickle\", \"rb\") as handle:\n",
    "        dict_of_individual_y_preds = pickle.load(handle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "all_participant_rs = []\n",
    "dims = dict_of_individual_ys.keys()\n",
    "for dim in dims:\n",
    "    df_ys      = dict_of_individual_ys[dim]\n",
    "#     df_ys      = dict_of_individual_ys[dim].iloc[:,:100]\n",
    "    df_y_preds = dict_of_individual_y_preds[dim]\n",
    "    dim_rs = []\n",
    "    for ys, y_preds in zip(df_ys.values, df_y_preds.values):\n",
    "        y_not_na_idx = ~np.isnan(ys.astype(float))\n",
    "\n",
    "        ys_not_na      = ys[y_not_na_idx]\n",
    "        y_preds_not_na = y_preds[y_not_na_idx]\n",
    "        \n",
    "        r, _ = pearsonr(ys_not_na, y_preds_not_na)\n",
    "        dim_rs.append(r)\n",
    "    all_participant_rs.append(dim_rs)\n",
    "all_participant_rs_df = pd.DataFrame(all_participant_rs, index=dims).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "all_participant_rs_df_stacked = all_participant_rs_df.stack().reset_index()\n",
    "all_participant_rs_df_stacked.columns = ['participant_number', 'Judgment Dimension', 'Correlation of Pred vs Actual']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "all_participant_rs_df_stacked.to_csv('results/individual-level_mapping_model_results.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,8))\n",
    "plt.xticks(rotation=45);\n",
    "sns.barplot(data=all_participant_rs_df_stacked, x='Judgment Dimension', y='Correlation of Pred vs Actual', color='gray');\n",
    "plt.savefig('figures/individual_level_mapping_pred_v_actual_correlation_barplots.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "participant_number               17.357143\n",
       "Correlation of Pred vs Actual     0.522605\n",
       "dtype: float64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_participant_rs_df_stacked.groupby('Judgment Dimension').mean().mean()"
   ]
  },
  {
   "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}