{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Table of Contents\n",
"\n",
"1. [Datasets](#dataset)
\n",
"2. [Total Imprisonment Rates](#2)
\n",
" 2a. [Visualization Pitfall](#3)
\n",
" 2b. [Population Dataset](#4)
\n",
" 2c. [Visualization Breakthrough, Observations, and interactive maps](#5)
\n",
"3. [State Imprisonment Distributions by Gender](#6)
\n",
" 3a. [Static Scatter Plots with Distributions](#7)
\n",
" 3b. [Interactive Scatter Plots](#8)
\n",
"4. [Summary](#9)
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Dataset: Taiwanese Credit Card Default Prediction \n",
"\n",
"https://archive.ics.uci.edu/ml/datasets/default+of+credit+card+clients\n",
"\n",
"this is where I expect the solutions to my issues will be: https://github.com/amueller/ml-training-advanced/blob/48162f50954a524ecb69b9d85de5233d886f8576/notebooks/extra%20-%20imbalanced%20datasets.ipynb\n",
"\n",
"https://scipy2017.scipy.org/ehome/220975/493423/"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {},
"outputs": [],
"source": [
"# General Utility\n",
"import os\n",
"import pandas as pd\n",
"import numpy as np\n",
"import seaborn as sns\n",
"# import pickle\n",
"from scipy import interp\n",
"import matplotlib.pyplot as plt\n",
"# import matplotlib.gridspec as gridspec\n",
"from matplotlib.collections import LineCollection\n",
"from mpl_toolkits.axes_grid1 import make_axes_locatable\n",
"from IPython.core.display import display, HTML\n",
"%matplotlib inline"
]
},
{
"cell_type": "code",
"execution_count": 187,
"metadata": {},
"outputs": [],
"source": [
"# Preprocessing\n",
"import missingno as msno\n",
"from sklearn.model_selection import train_test_split \n",
"from sklearn.preprocessing import RobustScaler\n",
"from imblearn.over_sampling import SMOTE \n",
"from imblearn.over_sampling import ADASYN\n",
"from imblearn.pipeline import Pipeline as imbPipeline\n",
"from imblearn.pipeline import make_pipeline as make_imb_pipeline\n",
"from sklearn.pipeline import Pipeline\n",
"from sklearn.pipeline import make_pipeline\n",
"from sklearn.preprocessing import StandardScaler\n",
"from sklearn.base import clone"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {},
"outputs": [],
"source": [
"# Model Building\n",
"from sklearn.model_selection import GridSearchCV\n",
"from sklearn.linear_model import LogisticRegression\n",
"from sklearn.neighbors import KNeighborsClassifier\n",
"from sklearn.ensemble import RandomForestClassifier\n",
"from sklearn.ensemble import BaggingClassifier\n",
"from sklearn.ensemble import AdaBoostClassifier\n",
"from sklearn.tree import DecisionTreeClassifier"
]
},
{
"cell_type": "code",
"execution_count": 190,
"metadata": {},
"outputs": [],
"source": [
"# Model Validation\n",
"from sklearn.metrics import accuracy_score\n",
"from sklearn.metrics import precision_score\n",
"from sklearn.metrics import precision_recall_curve\n",
"from sklearn.metrics import recall_score\n",
"from sklearn.metrics import log_loss\n",
"from sklearn.metrics import roc_auc_score\n",
"from sklearn.metrics import roc_curve\n",
"from sklearn.metrics import auc\n",
"from sklearn.metrics import confusion_matrix\n",
"from sklearn.metrics import classification_report\n",
"from sklearn.model_selection import cross_val_score\n",
"from sklearn.model_selection import StratifiedKFold\n",
"from sklearn.model_selection import RepeatedStratifiedKFold\n",
"from sklearn.metrics import precision_recall_curve"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"['bmh', 'classic', 'dark_background', 'fivethirtyeight', 'ggplot', 'grayscale', 'seaborn-bright', 'seaborn-colorblind', 'seaborn-dark-palette', 'seaborn-dark', 'seaborn-darkgrid', 'seaborn-deep', 'seaborn-muted', 'seaborn-notebook', 'seaborn-paper', 'seaborn-pastel', 'seaborn-poster', 'seaborn-talk', 'seaborn-ticks', 'seaborn-white', 'seaborn-whitegrid']\n"
]
}
],
"source": [
"print(plt.style.available)"
]
},
{
"cell_type": "code",
"execution_count": 64,
"metadata": {},
"outputs": [],
"source": [
"# Global\n",
"nb_seed = 1234\n",
"# plt.style.use('seaborn-darkgrid')\n",
"sns.set()"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
""
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"pd.options.display.max_columns = None\n",
"display(HTML(\"\"))\n",
"pd.set_option('display.float_format',lambda x: '%.4f' % x)\n",
"plt.rcParams['figure.figsize'] = 10,10"
]
},
{
"cell_type": "code",
"execution_count": 182,
"metadata": {},
"outputs": [],
"source": [
"def roc_plotter(fitted_clf, data_, labels_, ax=None):\n",
" \"\"\"\n",
" Formats and plots the ROC curve and colors the curve by its \n",
" threshold value at each point.\n",
" Args:\n",
" fitted_clf: A classifier that has already been fit to training data and labels\n",
" test_data: An array containing the corresponding testing data \"\"\"\n",
" if ax == None:\n",
" fig, ax = plt.subplots(figsize=(8, 8))\n",
" ax.plot([0, 1], [0, 1], linestyle='--', lw=2, color='r', label='Random', alpha=.8)\n",
" y_pred_prob = fitted_clf.predict_proba(data_)[:,1]\n",
" fpr, tpr, thresh = roc_curve(labels_, y_pred_prob) \n",
" mean_auc = auc(fpr, tpr)\n",
" lc = colorline(ax, fpr, tpr, thresh)\n",
" ax.plot(fpr, tpr, color='k',\n",
" label='Mean ROC (AUC = {:0.2f})'\n",
" .format(mean_auc), lw=1, alpha=0.5)\n",
" ax.legend(loc='best', fancybox=True, framealpha=0.4)\n",
" ax.set_title('ROC Curve', fontsize=14)\n",
" ax.set_ylabel('True Positive Rate', fontsize=14)\n",
" ax.set_xlabel('False Positive Rate', fontsize=14)\n",
" ax.set_xlim([0.0,1.0])\n",
" ax.set_ylim([0.0,1.0])\n",
" divider = make_axes_locatable(ax)\n",
" fig = ax.get_figure()\n",
" cax = divider.append_axes('right', size='5%', pad=0.05)\n",
" fig.colorbar(lc, cax=cax, orientation='vertical')\n",
" return ax"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"# Used the example at this address to encode threshold to color on the ROC curve # Used t \n",
"# https://nbviewer.jupyter.org/github/dpsanders/matplotlib-examples/blob/master/colorline.ipynb\n",
"def make_segments(x, y):\n",
" '''\n",
" Create list of line segments from x and y coordinates, in the correct format for LineCollection:\n",
" an array of the form numlines x (points per line) x 2 (x and y) array\n",
" '''\n",
" points = np.array([x, y]).T.reshape(-1, 1, 2)\n",
" segments = np.concatenate([points[:-1], points[1:]], axis=1)\n",
" return segments\n",
"\n",
"\n",
"def colorline(ax, x, y, z=None, cmap='coolwarm', \n",
" norm=plt.Normalize(0.0, 1.0), linewidth=10, alpha=0.9):\n",
" '''\n",
" Plot a colored line with coordinates x and y\n",
" Optionally specify colors in the array z\n",
" Optionally specify a colormap, a norm function and a line width\n",
" ''' \n",
" # Default colors equally spaced on [0,1]:\n",
" if z is None:\n",
" z = np.linspace(0.0, 1.0, len(x))\n",
" # Special case if a single number:\n",
" if not hasattr(z, \"__iter__\"): # to check for numerical input -- this is a hack\n",
" z = np.array([z])\n",
" z = np.asarray(z)\n",
" segments = make_segments(x, y)\n",
" lc = LineCollection(segments, array=z, cmap=cmap, norm=norm, linewidth=linewidth, alpha=alpha)\n",
" ax.add_collection(lc)\n",
" return lc"
]
},
{
"cell_type": "code",
"execution_count": 114,
"metadata": {},
"outputs": [],
"source": [
"g_cmap = sns.light_palette('seagreen', n_colors=10, as_cmap=True)\n",
"def confusion_mat_plotter(fitted_clf, X_train_, X_test_, y_train_, y_test_,\n",
" class_labels, cmap=g_cmap, ax=None):\n",
" \"\"\"\n",
" Generates plots of the confusion matrices for a classifier's performance\n",
" with both the training data (to check for overfitting) and the testing data.\n",
" Args:\n",
" fitted_clf: A classifier that has already been fit to training data and labels\n",
" X_train, ... Arrays containing the training/testing data/labels.\"\"\"\n",
" if ax == None:\n",
" fig, ax = plt.subplots(ncols=2, figsize=(13, 6.5))\n",
" \n",
" y_pred_train = fitted_clf.predict(X_train_)\n",
" y_pred_test = fitted_clf.predict(X_test_)\n",
" cfn_train = confusion_matrix(y_train_, y_pred_train)\n",
" cfn_test = confusion_matrix(y_test_, y_pred_test)\n",
" \n",
" sns.heatmap(cfn_train.T, square=True, annot=True, fmt='d', cbar=True, \n",
" xticklabels=class_labels, yticklabels=class_labels, ax=ax[0], cmap=g_cmap,\n",
" annot_kws={\"size\": 16})\n",
" ax[0].set_xlabel('True Class', fontsize=14)\n",
" ax[0].set_ylabel('Predicted Class', fontsize=14)\n",
" ax[0].set_title('Confusion Matrix (training data)', fontsize=14)\n",
" sns.heatmap(cfn_test.T, square=True, annot=True, fmt='d', cbar=True, \n",
" xticklabels=class_labels, yticklabels=class_labels, ax=ax[1], cmap=g_cmap,\n",
" annot_kws={\"size\": 16})\n",
" ax[1].set_xlabel('True Class', fontsize=14)\n",
" ax[1].set_ylabel('Predicted Class', fontsize=14)\n",
" ax[1].set_title('Confusion Matrix (testing data)', fontsize=14)\n",
" plt.tight_layout()"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [],
"source": [
"# avg_cfn = np.mean(cfns, axis=0).astype(int)\n",
"g_cmap = sns.light_palette('seagreen', n_colors=10, as_cmap=True)\n",
"def confusion_mat_plotter_single(cfn_matrix, class_labels, cmap=g_cmap, ax=None):\n",
" if ax == None:\n",
" fig, ax = plt.subplots(figsize=(6, 6))\n",
" sns.heatmap(cfn_matrix.T, square=True, annot=True, fmt='d', cbar=True, \n",
" xticklabels=class_labels, yticklabels=class_labels, ax=ax, cmap=g_cmap,\n",
" annot_kws={\"size\": 16})\n",
" ax.set_xlabel('True Class', fontsize=14)\n",
" ax.set_ylabel('Predicted Class', fontsize=14)\n",
" ax.set_title('Confusion Matrix', fontsize=14)"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"ID\n",
"1 1\n",
"2 1\n",
"3 0\n",
"4 0\n",
"5 0\n",
"Name: default_payment_next_month, dtype: int64"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" LIMIT_BAL | \n",
" SEX | \n",
" EDUCATION | \n",
" MARRIAGE | \n",
" AGE | \n",
" PAY_0 | \n",
" PAY_2 | \n",
" PAY_3 | \n",
" PAY_4 | \n",
" PAY_5 | \n",
" PAY_6 | \n",
" BILL_AMT1 | \n",
" BILL_AMT2 | \n",
" BILL_AMT3 | \n",
" BILL_AMT4 | \n",
" BILL_AMT5 | \n",
" BILL_AMT6 | \n",
" PAY_AMT1 | \n",
" PAY_AMT2 | \n",
" PAY_AMT3 | \n",
" PAY_AMT4 | \n",
" PAY_AMT5 | \n",
" PAY_AMT6 | \n",
"
\n",
" \n",
" | ID | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
"
\n",
" \n",
" \n",
" \n",
" | 1 | \n",
" 20000 | \n",
" 2 | \n",
" 2 | \n",
" 1 | \n",
" 24 | \n",
" 2 | \n",
" 2 | \n",
" -1 | \n",
" -1 | \n",
" -2 | \n",
" -2 | \n",
" 3913 | \n",
" 3102 | \n",
" 689 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 689 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
"
\n",
" \n",
" | 2 | \n",
" 120000 | \n",
" 2 | \n",
" 2 | \n",
" 2 | \n",
" 26 | \n",
" -1 | \n",
" 2 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 2 | \n",
" 2682 | \n",
" 1725 | \n",
" 2682 | \n",
" 3272 | \n",
" 3455 | \n",
" 3261 | \n",
" 0 | \n",
" 1000 | \n",
" 1000 | \n",
" 1000 | \n",
" 0 | \n",
" 2000 | \n",
"
\n",
" \n",
" | 3 | \n",
" 90000 | \n",
" 2 | \n",
" 2 | \n",
" 2 | \n",
" 34 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 29239 | \n",
" 14027 | \n",
" 13559 | \n",
" 14331 | \n",
" 14948 | \n",
" 15549 | \n",
" 1518 | \n",
" 1500 | \n",
" 1000 | \n",
" 1000 | \n",
" 1000 | \n",
" 5000 | \n",
"
\n",
" \n",
" | 4 | \n",
" 50000 | \n",
" 2 | \n",
" 2 | \n",
" 1 | \n",
" 37 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 46990 | \n",
" 48233 | \n",
" 49291 | \n",
" 28314 | \n",
" 28959 | \n",
" 29547 | \n",
" 2000 | \n",
" 2019 | \n",
" 1200 | \n",
" 1100 | \n",
" 1069 | \n",
" 1000 | \n",
"
\n",
" \n",
" | 5 | \n",
" 50000 | \n",
" 1 | \n",
" 2 | \n",
" 1 | \n",
" 57 | \n",
" -1 | \n",
" 0 | \n",
" -1 | \n",
" 0 | \n",
" 0 | \n",
" 0 | \n",
" 8617 | \n",
" 5670 | \n",
" 35835 | \n",
" 20940 | \n",
" 19146 | \n",
" 19131 | \n",
" 2000 | \n",
" 36681 | \n",
" 10000 | \n",
" 9000 | \n",
" 689 | \n",
" 679 | \n",
"
\n",
" \n",
"
\n",
"