Bankruptcy_in_Taiwan.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "163b83e6",
   "metadata": {},
   "source": [
    "<font size=\"+3\"><strong>5.5. Bankruptcy in Taiwan 🇹🇼</strong></font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b2a3a260",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import libraries here\n",
    "import gzip\n",
    "import json\n",
    "import pickle\n",
    "\n",
    "import ipywidgets as widgets\n",
    "import pandas as pd\n",
    "import wqet_grader\n",
    "from imblearn.over_sampling import RandomOverSampler\n",
    "from IPython.display import VimeoVideo\n",
    "from ipywidgets import interact\n",
    "from sklearn.ensemble import GradientBoostingClassifier\n",
    "from sklearn.impute import SimpleImputer\n",
    "from sklearn.metrics import (\n",
    "    ConfusionMatrixDisplay,\n",
    "    classification_report,\n",
    "    confusion_matrix,\n",
    ")\n",
    "from sklearn.model_selection import GridSearchCV, train_test_split\n",
    "from sklearn.pipeline import make_pipeline\n",
    "from teaching_tools.widgets import ConfusionMatrixWidget\n",
    "\n",
    "wqet_grader.init(\"Project 5 Assessment\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9c8187c0",
   "metadata": {},
   "source": [
    "# Prepare Data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "97c505f0",
   "metadata": {},
   "source": [
    "## Import"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b11461cd",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load data file\n",
    "with gzip.open(\"taiwan-bankruptcy-data.json.gz\", \"r\") as f:\n",
    "        taiwan_data = json.load(f)\n",
    "print(type(taiwan_data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f6e6f6f5",
   "metadata": {},
   "outputs": [],
   "source": [
    "taiwan_data_keys = taiwan_data.keys()\n",
    "print(taiwan_data_keys)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d4baa810",
   "metadata": {},
   "outputs": [],
   "source": [
    "n_companies = len(taiwan_data[\"observations\"])\n",
    "print(n_companies)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "855ce08b",
   "metadata": {},
   "outputs": [],
   "source": [
    "n_features = len(taiwan_data[\"schema\"]['fields'])\n",
    "print(n_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "441254e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create wrangle function\n",
    "def wrangle(filename):\n",
    "    \n",
    "    # Open compressed file, load into dictionary\n",
    "    with gzip.open(filename, \"r\") as f:\n",
    "        data = json.load(f)\n",
    "\n",
    "    # Load dictionary into DataFrame, set index\n",
    "    df = pd.DataFrame().from_dict(data[\"observations\"]).set_index(\"id\")\n",
    "\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bc7ce4fd",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = wrangle(\"data/taiwan-bankruptcy-data.json.gz\")\n",
    "print(\"df shape:\", df.shape)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c671f226",
   "metadata": {},
   "outputs": [],
   "source": [
    "df.isna().any().shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9a739637",
   "metadata": {},
   "outputs": [],
   "source": [
    "nans_by_col = df.isna().any()\n",
    "print(\"nans_by_col shape:\", nans_by_col.shape)\n",
    "nans_by_col.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c713169c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Plot class balance\n",
    "df[\"bankrupt\"].value_counts(normalize=True).plot(kind=\"bar\");\n",
    "plt.xlabel(\"Bankrupt\")\n",
    "plt.ylabel(\"Frequency\")\n",
    "plt.title(\"Class Balance\")\n",
    "# Don't delete the code below 👇\n",
    "plt.savefig(\"images/5-5-7.png\", dpi=150)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6c99a630",
   "metadata": {},
   "outputs": [],
   "source": [
    "target = \"bankrupt\"\n",
    "X = df.drop(columns=target)\n",
    "y = df[target]\n",
    "print(\"X shape:\", X.shape)\n",
    "print(\"y shape:\", y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "88e63d46",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n",
    "print(\"X_train shape:\", X_train.shape)\n",
    "print(\"y_train shape:\", y_train.shape)\n",
    "print(\"X_test shape:\", X_test.shape)\n",
    "print(\"y_test shape:\", y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "90af1d0a",
   "metadata": {},
   "outputs": [],
   "source": [
    "over_sampler = RandomOverSampler(random_state=42)\n",
    "X_train_over, y_train_over = over_sampler.fit_resample(X_train, y_train)\n",
    "print(\"X_train_over shape:\", X_train_over.shape)\n",
    "X_train_over.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8398ddd7",
   "metadata": {},
   "source": [
    "# Build Model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e561c7d5",
   "metadata": {},
   "source": [
    "## Iterate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fe7f8962",
   "metadata": {},
   "outputs": [],
   "source": [
    "clf = GradientBoostingClassifier()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "73c0827f",
   "metadata": {},
   "outputs": [],
   "source": [
    "cv_scores = cross_val_score(clf, X_train_over, y_train_over, cv=5, n_jobs=-1)\n",
    "print(cv_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "06f312db",
   "metadata": {},
   "outputs": [],
   "source": [
    "params = {\n",
    "    \"gradientboostingclassifier__n_estimators\": range(20, 31, 5),\n",
    "    \"gradientboostingclassifier__max_depth\": range(2, 5,)\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "27f60522",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = GridSearchCV(\n",
    "    clf,\n",
    "    param_grid = params,\n",
    "    cv = 5,\n",
    "    n_jobs = -1,\n",
    "    verbose=1\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f8548b67",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.fit(X_train_over, y_train_over)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5ecf043f",
   "metadata": {},
   "outputs": [],
   "source": [
    "cv_results = pd.DataFrame(model.cv_results_)\n",
    "cv_results.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a0c3803e",
   "metadata": {},
   "outputs": [],
   "source": [
    "best_params = model.best_params_\n",
    "print(best_params)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "75136a52",
   "metadata": {},
   "source": [
    "## Evaluate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a9da4bef",
   "metadata": {},
   "outputs": [],
   "source": [
    "acc_train = model.score(X_train, y_train)\n",
    "acc_test = model.score(X_test, y_test)\n",
    "\n",
    "print(\"Model Training Accuracy:\", round(acc_train, 4))\n",
    "print(\"Model Test Accuracy:\", round(acc_test, 4))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2764f528",
   "metadata": {},
   "outputs": [],
   "source": [
    "ConfusionMatrixDisplay.from_estimator(model, X_test, y_test)\n",
    "# Don't delete the code below 👇\n",
    "plt.savefig(\"images/5-5-16.png\", dpi=150)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a105c700",
   "metadata": {},
   "outputs": [],
   "source": [
    "class_report = classification_report(y_test, model.predict(X_test))\n",
    "print(class_report)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bca218db",
   "metadata": {},
   "source": [
    "# Communicate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bd49783d",
   "metadata": {},
   "outputs": [],
   "source": [
    "features = X_train_over.columns\n",
    "importances = model.best_estimator_.feature_importances_\n",
    "feat_imp = pd.Series(importances, index=features).sort_values()\n",
    "feat_imp.tail(10).plot(kind=\"barh\")\n",
    "plt.xlabel(\"Gini Importance\")\n",
    "plt.ylabel(\"Feature\")\n",
    "plt.title(\"Feature Importance\");\n",
    "# Don't delete the code below 👇\n",
    "plt.savefig(\"images/5-5-17.png\", dpi=150)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "df4cceb2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save model\n",
    "with open(\"model-5-5.pkl\", \"wb\") as f:\n",
    "    pickle.dump(model, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "972ad897",
   "metadata": {},
   "outputs": [],
   "source": [
    "def wrangle(filename):\n",
    "    \n",
    "    # Open compressed file, load into dictionary\n",
    "    with gzip.open(filename, \"r\") as f:\n",
    "        data = json.load(f)\n",
    "\n",
    "    # Load dictionary into DataFrame, set index\n",
    "    df = pd.DataFrame().from_dict(data[\"observations\"]).set_index(\"id\")\n",
    "\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "14b730ab",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add make_predictions function from lesson 5.3\n",
    "def make_predictions(data_filepath, model_filepath):\n",
    "    # Wrangle JSON file\n",
    "    X_test = wrangle(data_filepath)\n",
    "    # Load model\n",
    "    with open(model_filepath, \"rb\")as f:\n",
    "        model = pickle.load(f)\n",
    "    # Generate predictions\n",
    "    y_test_pred = model.predict(X_test)\n",
    "    # Put predictions into Series with name \"bankrupt\", and same index as X_test\n",
    "    y_test_pred = pd.Series(y_test_pred, index=X_test.index, name=\"bankrupt\")\n",
    "    return y_test_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e8d7cb57",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import your module\n",
    "#from my_predictor_assignment import make_predictions\n",
    "\n",
    "# Generate predictions\n",
    "y_test_pred = make_predictions(\n",
    "    data_filepath=\"taiwan-bankruptcy-data-test-features.json.gz\",\n",
    "    model_filepath=\"model-5-5.pkl\",\n",
    ")\n",
    "\n",
    "print(\"predictions shape:\", y_test_pred.shape)\n",
    "y_test_pred.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5f8fb052",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}