Interactive Skills Laboratory

# Customer Segmentation & RFM Analysis
import pandas as pd
import numpy as np
from sklearn.preprocessing import StandardScaler

def generate_sample_data(num_customers=1000):
    """Generate customer data for RFM analysis"""
    np.random.seed(42)

    # Generate RFM data
    recency = np.random.randint(2, 46, size=num_customers)
    frequency = np.random.randint(3, 21, size=num_customers)
    avg_order_value = np.full(num_customers, 100)
    monetary = frequency * avg_order_value

    # Create DataFrame
    df = pd.DataFrame({
        'recency': recency,
        'frequency': frequency,
        'monetary': monetary,
        'orders': frequency,
        'avg_order_value': avg_order_value
    })

    df['total_spent'] = df['orders'] * df['avg_order_value']

    # Scale features
    scaler = StandardScaler()
    scaler.fit_transform(df[['recency', 'frequency', 'monetary']])

    # Calculate results
    return {
        'high_value': len(df[df['total_spent'] > df['total_spent'].quantile(0.8)]),
        'avg_clv': round(df['total_spent'].mean(), 2),
        'retention_rate': round((df['recency'] < 30).mean() * 100, 2)
    }

# Run analysis on 1000 customers
result = generate_sample_data(1000)
print(f"High-value customers: {result['high_value']}")
print(f"Average CLV: ${result['avg_clv']:.2f}")
print(f"Retention rate: {result['retention_rate']:.1f}%")

# Statistical Analysis in Python
import numpy as np
from scipy import stats

# A/B Test Analysis Function
def analyze_ab_test(control, treatment):
    """Perform A/B test analysis with t-test and effect size"""
    control = np.array(control)
    treatment = np.array(treatment)

    # Calculate basic statistics
    control_mean = np.mean(control)
    treatment_mean = np.mean(treatment)

    # Perform independent samples t-test
    t_stat, p_value = stats.ttest_ind(treatment, control)

    # Calculate effect size (Cohen's d)
    n1, n2 = len(control), len(treatment)
    var1 = np.var(control, ddof=1)
    var2 = np.var(treatment, ddof=1)
    pooled_std = np.sqrt(((n1-1)*var1 + (n2-1)*var2) / (n1+n2-2))
    effect_size = (treatment_mean - control_mean) / pooled_std

    return {
        'control_mean': control_mean,
        'treatment_mean': treatment_mean,
        'p_value': p_value,
        'effect_size': effect_size,
        'significant': p_value < 0.05
    }

# Sample data
control_group = [2.3, 2.1, 2.4, 2.2, 2.5, 2.0, 2.3]
treatment_group = [2.8, 2.9, 2.7, 3.0, 2.6, 2.8, 2.9]

# Run analysis
results = analyze_ab_test(control_group, treatment_group)
lift = (results['treatment_mean'] / results['control_mean'] - 1) * 100
print(f"Lift: {lift:.1f}%")
print(f"P-value: {results['p_value']:.4f}")
print(f"Significant: {results['significant']}")