shap_plots.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**SHAP Explainability Plots** \\\n",
    "_Author: Joaquín Torres Bravo_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import shap\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.patches as mpatches\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "ename": "FileNotFoundError",
     "evalue": "[Errno 2] No such file or directory: '../EDA/output/feature_names/all_features.npy'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mFileNotFoundError\u001b[0m                         Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[3], line 2\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[38;5;66;03m# Retrieve attribute names in order\u001b[39;00m\n\u001b[0;32m----> 2\u001b[0m attribute_names \u001b[38;5;241m=\u001b[39m attribute_names \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mlist\u001b[39m(\u001b[43mnp\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mload\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43m../EDA/output/feature_names/all_features.npy\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mallow_pickle\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mTrue\u001b[39;49;00m\u001b[43m)\u001b[49m)\n\u001b[1;32m      4\u001b[0m \u001b[38;5;66;03m# Load test data\u001b[39;00m\n\u001b[1;32m      5\u001b[0m X_test_pre \u001b[38;5;241m=\u001b[39m np\u001b[38;5;241m.\u001b[39mload(\u001b[38;5;124m'\u001b[39m\u001b[38;5;124m../gen_train_data/output/pre/X_test_pre.npy\u001b[39m\u001b[38;5;124m'\u001b[39m, allow_pickle\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m)\n",
      "File \u001b[0;32m~/covid_analysis/.venv/lib/python3.9/site-packages/numpy/lib/npyio.py:427\u001b[0m, in \u001b[0;36mload\u001b[0;34m(file, mmap_mode, allow_pickle, fix_imports, encoding, max_header_size)\u001b[0m\n\u001b[1;32m    425\u001b[0m     own_fid \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mFalse\u001b[39;00m\n\u001b[1;32m    426\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m--> 427\u001b[0m     fid \u001b[38;5;241m=\u001b[39m stack\u001b[38;5;241m.\u001b[39menter_context(\u001b[38;5;28;43mopen\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43mos_fspath\u001b[49m\u001b[43m(\u001b[49m\u001b[43mfile\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mrb\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m)\u001b[49m)\n\u001b[1;32m    428\u001b[0m     own_fid \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mTrue\u001b[39;00m\n\u001b[1;32m    430\u001b[0m \u001b[38;5;66;03m# Code to distinguish from NumPy binary files and pickles.\u001b[39;00m\n",
      "\u001b[0;31mFileNotFoundError\u001b[0m: [Errno 2] No such file or directory: '../EDA/output/feature_names/all_features.npy'"
     ]
    }
   ],
   "source": [
    "# Retrieve attribute names in order\n",
    "attribute_names = attribute_names = list(np.load('../EDA/output/feature_names/all_features.npy', allow_pickle=True))\n",
    "\n",
    "# Load test data\n",
    "X_test_pre = np.load('../gen_train_data/output/pre/X_test_pre.npy', allow_pickle=True)\n",
    "y_test_pre = np.load('../gen_train_data/output/pre/y_test_pre.npy', allow_pickle=True)\n",
    "X_test_post = np.load('../gen_train_data/output/post/X_test_post.npy', allow_pickle=True)\n",
    "y_test_post = np.load('../gen_train_data/output/post/y_test_post.npy', allow_pickle=True)\n",
    "\n",
    "# Type conversion needed    \n",
    "data_dic = {\n",
    "    \"X_test_pre\": pd.DataFrame(X_test_pre, columns=attribute_names).convert_dtypes(),\n",
    "    \"y_test_pre\": y_test_pre,\n",
    "    \"X_test_post\": pd.DataFrame(X_test_post, columns=attribute_names).convert_dtypes(),\n",
    "    \"y_test_post\": y_test_post,\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "method_names = {\n",
    "    0: \"ORIG\",\n",
    "    1: \"ORIG_CW\",\n",
    "    2: \"OVER\",\n",
    "    3: \"UNDER\"\n",
    "}\n",
    "model_choices = {\n",
    "    \"ORIG\": \"XGB\",\n",
    "    \"ORIG_CW\": \"RF\",\n",
    "    \"OVER\": \"XGB\",\n",
    "    \"UNDER\": \"XGB\"\n",
    "}\n",
    "\n",
    "# Load names of social and individual attributes\n",
    "soc_var_names = np.load('../EDA/output/feature_names/social_factors.npy', allow_pickle=True)\n",
    "ind_var_names = np.load('../EDA/output/feature_names/individual_factors.npy', allow_pickle=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### SHAP Plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "method_name = 'OVER'\n",
    "\n",
    "plt.figure(figsize=(35, 75))\n",
    "for i, group in enumerate(['pre', 'post']):\n",
    "            X_test = data_dic['X_test_' + group]\n",
    "            y_test = data_dic['y_test_' + group]\n",
    "            model_name = model_choices[method_name]\n",
    "            shap_vals = np.load(f'./output/shap_values/{group}_{method_name}.npy')\n",
    "            ax = plt.subplot(2,1,i+1) # 2 rows (pre - post) 1 column\n",
    "            # show = False to modify plot before showing\n",
    "            shap.summary_plot(shap_vals, X_test, max_display=len(attribute_names), show=False)\n",
    "            plt.title(group.upper(), fontsize = 12, fontweight='bold')\n",
    "            plt.xlabel('SHAP Value')\n",
    "            plt.xlim(-3,5)\n",
    "            used_colors = {'purple': 'Social factor', 'green': 'Individual factor'}\n",
    "            # Modify color of attributes\n",
    "            for label in ax.get_yticklabels():\n",
    "                label_text = label.get_text()  # Get the text of the label\n",
    "                label.set_fontsize(8)\n",
    "                if label_text in soc_var_names:\n",
    "                        label.set_color('purple')\n",
    "                else:\n",
    "                        label.set_color('green')\n",
    "                # Create custom legend for each subplot\n",
    "                handles = [mpatches.Patch(color=color, label=label) for color, label in used_colors.items()]\n",
    "                ax.legend(handles=handles, loc='lower right', fontsize=8)\n",
    "            \n",
    "plt.suptitle(f'SHAP Summary Plots PRE vs POST - Pipeline: Oversampling - Model: {model_name}\\n\\n')\n",
    "plt.subplots_adjust(wspace=1)\n",
    "plt.tight_layout()\n",
    "plt.savefig(f'./output/plots/shap_summary/{method_name}_{model_name}.svg', format='svg', dpi=1250)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'plt' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[1], line 2\u001b[0m\n\u001b[1;32m      1\u001b[0m method_name \u001b[38;5;241m=\u001b[39m \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mORIG_CW\u001b[39m\u001b[38;5;124m'\u001b[39m\n\u001b[0;32m----> 2\u001b[0m \u001b[43mplt\u001b[49m\u001b[38;5;241m.\u001b[39mfigure(figsize\u001b[38;5;241m=\u001b[39m(\u001b[38;5;241m35\u001b[39m, \u001b[38;5;241m75\u001b[39m))\n\u001b[1;32m      3\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m i, group \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28menumerate\u001b[39m([\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mpre\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mpost\u001b[39m\u001b[38;5;124m'\u001b[39m]):\n\u001b[1;32m      4\u001b[0m             X_test \u001b[38;5;241m=\u001b[39m data_dic[\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mX_test_\u001b[39m\u001b[38;5;124m'\u001b[39m \u001b[38;5;241m+\u001b[39m group]\n",
      "\u001b[0;31mNameError\u001b[0m: name 'plt' is not defined"
     ]
    }
   ],
   "source": [
    "method_name = 'ORIG_CW'\n",
    "plt.figure(figsize=(35, 75))\n",
    "for i, group in enumerate(['pre', 'post']):\n",
    "            X_test = data_dic['X_test_' + group]\n",
    "            y_test = data_dic['y_test_' + group]\n",
    "            model_name = model_choices[method_name]\n",
    "            shap_vals = np.load(f'./output/shap_values/{group}_{method_name}.npy')\n",
    "            shap_vals = shap_vals[:,:,1] # Select shap values for positive class\n",
    "            ax = plt.subplot(2,1,i+1)\n",
    "            shap.summary_plot(shap_vals, X_test, max_display=len(attribute_names), show=False)\n",
    "            plt.title(group.upper(), fontsize = 12, fontweight='bold')\n",
    "            plt.xlabel('SHAP Value')\n",
    "            plt.xlim(-0.5,0.5)\n",
    "            used_colors = {'purple': 'Social factor', 'green': 'Individual factor'}\n",
    "            for label in ax.get_yticklabels():\n",
    "                label_text = label.get_text()\n",
    "                label.set_fontsize(8)\n",
    "                if label_text in soc_var_names:\n",
    "                        label.set_color('purple')\n",
    "                else:\n",
    "                        label.set_color('green')\n",
    "                handles = [mpatches.Patch(color=color, label=label) for color, label in used_colors.items()]\n",
    "                ax.legend(handles=handles, loc='lower right', fontsize=8)\n",
    "\n",
    "plt.suptitle(f'SHAP Summary Plots PRE vs POST - Pipeline: Original with Class Weight - Model: {model_name}\\n\\n')\n",
    "plt.subplots_adjust(wspace=1)\n",
    "plt.tight_layout()\n",
    "plt.savefig(f'./output/plots/shap_summary/{method_name}_{model_name}.svg', format='svg', dpi=1250)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### SHAP Interaction Plots"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**IMPORTANT NOTE**\n",
    "For the code to work as intended, the SHAP source code had to be modified:\n",
    "* **File**: shap/plots/_beeswarm.py\n",
    "* **Modification**: .venv/lib/python3.9/site-packages/shap/plots/_beeswarm.py \\\n",
    "\n",
    "sort_inds = np.argsort(-np.abs(shap_values.sum(1)).sum(0)) \\\n",
    "**replaced by** \\\n",
    "sort_inds = np.arange(39)\n",
    "\n",
    "The idea is to display the features in the original order instead of sorting them according to their absolute SHAP impact, as the library originally does\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Had to modify beeswarm.py (line 591)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "method_name = 'ORIG_CW'\n",
    "group = 'pre'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = data_dic['X_test_' + group]\n",
    "y_test = data_dic['y_test_' + group]\n",
    "model_name = model_choices[method_name]\n",
    "\n",
    "shap_inter_vals = np.load(f'./output/shap_inter_values/{group}_{method_name}.npy')\n",
    "if method_name == 'ORIG_CW':\n",
    "      shap_inter_vals = shap_inter_vals[:,:,:,1] # Take info about positive class\n",
    "\n",
    "num_instances = shap_inter_vals.shape[0]  # Dynamically get the number of instances\n",
    "num_features = shap_inter_vals.shape[1]  # Assuming the number of features is the second dimension size\n",
    "# Loop over each instance and set the diagonal and lower triangle of each 39x39 matrix to NaN\n",
    "for i in range(num_instances):\n",
    "    # Mask the diagonal\n",
    "    np.fill_diagonal(shap_inter_vals[i], np.nan)\n",
    "    # Get indices for the lower triangle, excluding the diagonal\n",
    "    lower_triangle_indices = np.tril_indices(num_features, -1)  # -1 excludes the diagonal\n",
    "    # Set the lower triangle to NaN\n",
    "    shap_inter_vals[i][lower_triangle_indices] = np.nan\n",
    "\n",
    "plt.figure(figsize=(10,10))\n",
    "shap.summary_plot(shap_inter_vals, X_test, show=False, sort=False)\n",
    "fig=plt.gcf()\n",
    "attr_names = []\n",
    "used_colors = {'purple': 'Social factor', 'green': 'Individual factor'}\n",
    "# Iterate over all axes in the figure\n",
    "for ax in fig.get_axes():\n",
    "    # Customize the y-axis tick labels\n",
    "    for label in ax.get_yticklabels():\n",
    "        label_text = label.get_text()  # Get the text of the label\n",
    "        attr_names.append(label_text)\n",
    "        label.set_fontsize(12)\n",
    "        if label_text in soc_var_names:\n",
    "                label.set_color('purple')\n",
    "        else:\n",
    "                label.set_color('green')\n",
    "\n",
    "# Assuming the top labels are treated as titles, let's try to modify them\n",
    "total_axes = len(fig.axes)\n",
    "for i, ax in enumerate(fig.axes):\n",
    "        reverse_index = total_axes - 1 - i\n",
    "        title = attr_names[reverse_index]\n",
    "        ax.set_title(title, color='purple' if title in soc_var_names else 'green', fontsize=12, rotation=90)\n",
    "        if method_name == 'ORIG_CW':\n",
    "              ax.set_xlim(-0.15, 0.15) # Use same scale for pre and post\n",
    "        elif method_name == 'OVER':\n",
    "              ax.set_xlim(-2,2)\n",
    "\n",
    "# Create a single general legend for the whole figure\n",
    "handles = [mpatches.Patch(color=color, label=label) for color, label in used_colors.items()]\n",
    "fig.legend(handles=handles, loc='lower right', fontsize=12)\n",
    "\n",
    "plt.suptitle(f'Simplified Example SHAP Summary Interaction Plot\\n', fontsize=15, fontweight='bold', x=0.5, y=0.95, ha='center')\n",
    "plt.tight_layout()\n",
    "plt.savefig(f'./output/plots/shap_inter_summary/example.svg', format='svg', dpi=600)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### SHAP Interaction Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "method_name = 'OVER'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Normalization function\n",
    "def max_absolute_scale(matrix):\n",
    "    return matrix / np.max(np.abs(matrix))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load array of shap inter matrices for pre and post for the chosen method\n",
    "shap_inter_vals_pre = np.load(f'./output/shap_inter_values/pre_{method_name}.npy')\n",
    "shap_inter_vals_post = np.load(f'./output/shap_inter_values/post_{method_name}.npy')\n",
    "if method_name == 'ORIG_CW':\n",
    "    shap_inter_vals_pre = shap_inter_vals_pre[:,:,:,1]\n",
    "    shap_inter_vals_post = shap_inter_vals_post[:,:,:,1]\n",
    "\n",
    "# Normalize each matrix in each of the two arrays\n",
    "norm_shap_inter_vals_pre = np.array([max_absolute_scale(matrix) for matrix in shap_inter_vals_pre])\n",
    "norm_shap_inter_vals_post = np.array([max_absolute_scale(matrix) for matrix in shap_inter_vals_post])\n",
    "\n",
    "# Aggregate matrices in each group by calculating the mean\n",
    "agg_shap_inter_vals_pre = np.mean(norm_shap_inter_vals_pre, axis=0)\n",
    "agg_shap_inter_vals_post = np.mean(norm_shap_inter_vals_post, axis=0)\n",
    "\n",
    "# Compute the difference between the aggregated matrices and take the absolute value\n",
    "dist_matrix = np.abs(agg_shap_inter_vals_post - agg_shap_inter_vals_pre)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_inter_heatmap(matrix, feature_names, soc_var_names, method_name, group):\n",
    "    # Create a mask for the upper triangle\n",
    "    mask = np.triu(np.ones_like(matrix, dtype=bool))\n",
    "\n",
    "    # Create the heatmap using Seaborn\n",
    "    plt.figure(figsize=(12, 12))\n",
    "    ax = sns.heatmap(matrix, mask=mask, cmap='coolwarm', center=0, annot=False, cbar=True,\n",
    "                xticklabels=feature_names, yticklabels=feature_names)\n",
    "    \n",
    "    for tick_label in ax.get_yticklabels():\n",
    "        if tick_label.get_text() in soc_var_names:\n",
    "            tick_label.set_color('purple')  # Specific social variables\n",
    "        else:\n",
    "            tick_label.set_color('green')  # Other variables\n",
    "    \n",
    "    for tick_label in ax.get_xticklabels():\n",
    "        if tick_label.get_text() in soc_var_names:\n",
    "            tick_label.set_color('purple')  # Specific social variables\n",
    "        else:\n",
    "            tick_label.set_color('green')  # Other variables\n",
    "        \n",
    "    plt.title(f'Mean Interaction Matrix - Pipeline: {method_name} - Group: {group}\\n', fontdict={'fontstyle': 'normal', 'weight': 'bold'})\n",
    "    # Create a custom legend\n",
    "    purple_patch = mpatches.Patch(color='purple', label='Social factor')\n",
    "    green_patch = mpatches.Patch(color='green', label='Individual factor')\n",
    "    plt.legend(handles=[purple_patch, green_patch], loc='upper right')\n",
    "    # Add a title to the color bar\n",
    "    cbar = ax.collections[0].colorbar\n",
    "    cbar.set_label('Normalized SHAP Interaction Value', labelpad=15, rotation=270, verticalalignment='bottom')\n",
    "    \n",
    "    plt.savefig(f'./output/plots/heatmaps_interactions/{method_name}_{group}.svg', format='svg', dpi=600)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plot_inter_heatmap(agg_shap_inter_vals_post, attribute_names, soc_var_names, method_name, 'POST')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_distance_heatmap(matrix, feature_names, soc_var_names, method_name):\n",
    "    # Create a mask for the upper triangle\n",
    "    mask = np.triu(np.ones_like(matrix, dtype=bool))\n",
    "\n",
    "    # Create the heatmap using Seaborn\n",
    "    plt.figure(figsize=(12, 12))\n",
    "    ax = sns.heatmap(matrix, mask=mask, cmap=sns.color_palette(\"light:r\", as_cmap=True), annot=False, cbar=True,\n",
    "                xticklabels=feature_names, yticklabels=feature_names)\n",
    "\n",
    "    for tick_label in ax.get_yticklabels():\n",
    "        if tick_label.get_text() in soc_var_names:\n",
    "            tick_label.set_color('purple')  # Specific social variables\n",
    "        else:\n",
    "            tick_label.set_color('green')  # Other variables\n",
    "    \n",
    "    for tick_label in ax.get_xticklabels():\n",
    "        if tick_label.get_text() in soc_var_names:\n",
    "            tick_label.set_color('purple')  # Specific social variables\n",
    "        else:\n",
    "            tick_label.set_color('green')  # Other variables\n",
    "        \n",
    "    plt.title(f'Distance Interaction Matrix between PRE and POST - Pipeline: {method_name}\\n', fontdict={'fontstyle': 'normal', 'weight': 'bold'})\n",
    "    # Create a custom legend\n",
    "    purple_patch = mpatches.Patch(color='purple', label='Social factor')\n",
    "    green_patch = mpatches.Patch(color='green', label='Individual factor')\n",
    "    plt.legend(handles=[purple_patch, green_patch], loc='upper right')\n",
    "    # Add a title to the color bar\n",
    "    cbar = ax.collections[0].colorbar\n",
    "    cbar.set_label('Abs (PRE - POST)', labelpad=15, rotation=270, verticalalignment='bottom')\n",
    "\n",
    "    plt.savefig(f'./output/plots/heatmaps_interactions/DIST_{method_name}.svg', format='svg', dpi=600)\n",
    "    \n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plot_distance_heatmap(dist_matrix, attribute_names, soc_var_names, method_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define tolerance\n",
    "tolerance = np.median(dist_matrix)\n",
    "# Create a DataFrame to hold the interactions\n",
    "interactions = []\n",
    "\n",
    "# Iterate over the matrix to extract interactions above the tolerance\n",
    "for i in range(1, dist_matrix.shape[0]):\n",
    "    for j in range(i):  # Lower triangle exclduing diagonal\n",
    "        if dist_matrix[i, j] > tolerance:\n",
    "            var1 = attribute_names[i]\n",
    "            var2 = attribute_names[j]\n",
    "            if var1 in soc_var_names and var2 in soc_var_names:\n",
    "                inter_type = 'Social'\n",
    "            elif var1 in ind_var_names and var2 in ind_var_names:\n",
    "                inter_type = 'Individual'\n",
    "            else:\n",
    "                inter_type = 'Mixed'\n",
    "            interactions.append({\n",
    "                'Variable 1': var1, \n",
    "                'Variable 2': var2,\n",
    "                'SHAP Inter Variation PRE-POST': dist_matrix[i, j],\n",
    "                'Interaction Type': inter_type\n",
    "            })\n",
    "\n",
    "# Convert the list of dictionaries to a DataFrame\n",
    "interactions_df = pd.DataFrame(interactions)\n",
    "\n",
    "# Sort the DataFrame by 'Interaction Strength' in descending order\n",
    "sorted_interactions_df = interactions_df.sort_values(by='SHAP Inter Variation PRE-POST', ascending=False)\n",
    "\n",
    "# Export to Excel\n",
    "sorted_interactions_df.to_excel(f'./output/inter_variation_{method_name}.xlsx', index=False)\n",
    "\n",
    "print(\"Excel file has been created.\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.9.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}