Classification Modeling: Customer Churn Prediction with XGBoost

Acquiring a new customer costs 5x to 7x more than retaining an existing one. Customer Churn Prediction is the quintessential classification problem in enterprise data science. In this class, we build an end-to-end churn prediction pipeline using Logistic Regression, Random Forests, and XGBoost.

Algorithm Comparison: Linear vs. Tree Ensembles

End-to-End Customer Churn Model with XGBoost

import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report, roc_auc_score
from xgboost import XGBClassifier

# Generate synthetic subscription customer records
np.random.seed(42)
n = 2000
df = pd.DataFrame({
    'tenure_months': np.random.randint(1, 60, n),
    'monthly_charges': np.random.uniform(20, 120, n),
    'support_tickets': np.random.poisson(lam=2, size=n),
    'contract_annual': np.random.choice([0, 1], n, p=[0.7, 0.3])
})
# Churn probability increases with high tickets, low tenure, and month-to-month contracts
logits = (df['support_tickets'] * 0.4) - (df['tenure_months'] * 0.05) - (df['contract_annual'] * 1.2)
probs = 1 / (1 + np.exp(-logits))
df['churned'] = (probs > 0.50).astype(int)

X = df.drop(columns=['churned'])
y = df['churned']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, stratify=y, random_state=42)

# Train XGBoost Classifier
xgb_model = XGBClassifier(
    n_estimators=150,
    max_depth=4,
    learning_rate=0.08,
    subsample=0.8,
    colsample_bytree=0.8,
    eval_metric='logloss',
    random_state=42
)
xgb_model.fit(X_train, y_train)

# Predict probabilities
y_prob = xgb_model.predict_proba(X_test)[:, 1]
y_pred = (y_prob >= 0.35).astype(int)  # Business-optimized threshold

print(f"ROC-AUC Score: {roc_auc_score(y_test, y_prob):.4f}")
print("\n--- Classification Report (Threshold = 0.35) ---\n")
print(classification_report(y_test, y_pred, target_names=['Retained', 'Churned']))