┌─────────────────────────────────────────────────────────────────────────────┐
│                    AI-POWERED CREDIT SCORING                              │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                             │
│  ┌──────────────────────────────────────────────────────────────────────┐   │
│  │                    DATA INGESTION                                   │   │
│  │  (Traditional data + Alternative data + Real-time data)             │   │
│  └──────────────────────────────────────────────────────────────────────┘   │
│                                    │                                        │
│                                    v                                        │
│  ┌──────────────────────────────────────────────────────────────────────┐   │
│  │                    DATA PROCESSING                                  │   │
│  │  (Cleaning, normalisation, feature engineering)                     │   │
│  └──────────────────────────────────────────────────────────────────────┘   │
│                                    │                                        │
│                                    v                                        │
│  ┌──────────────────────────────────────────────────────────────────────┐   │
│  │                    MODEL TRAINING                                   │   │
│  │  (XGBoost, deep learning, ensemble models)                         │   │
│  └──────────────────────────────────────────────────────────────────────┘   │
│                                    │                                        │
│                                    v                                        │
│  ┌──────────────────────────────────────────────────────────────────────┐   │
│  │                    MODEL VALIDATION                                 │   │
│  │  (AUC, KS, calibration, fairness)                                  │   │
│  └──────────────────────────────────────────────────────────────────────┘   │
│                                    │                                        │
│                                    v                                        │
│  ┌──────────────────────────────────────────────────────────────────────┐   │
│  │                    REAL-TIME SCORING                                │   │
│  │  (Instant credit decisions)                                         │   │
│  └──────────────────────────────────────────────────────────────────────┘   │
│                                                                             │
└─────────────────────────────────────────────────────────────────────────────┘

┌─────────────────────────────────────────────────────────────────────────────┐
│                    DIGITAL LOAN ORIGINATION                               │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                             │
│  ┌─────────────┐    ┌─────────────┐    ┌─────────────┐                    │
│  │  Application │ → │  Data       │ → │  Credit     │                    │
│  │  (Mobile/Web)│    │  Verification│    │  Decision   │                    │
│  └─────────────┘    └─────────────┘    └─────────────┘                    │
│        │                  │                  │                             │
│        v                  v                  v                             │
│  ┌─────────────┐    ┌─────────────┐    ┌─────────────┐                    │
│  │  Offer      │ ← │  Approval   │ ← │  Pricing    │                    │
│  │  (Terms)    │    │  (Conditional)│    │  (Risk-based)│                  │
│  └─────────────┘    └─────────────┘    └─────────────┘                    │
│        │                  │                  │                             │
│        v                  v                  v                             │
│  ┌─────────────┐    ┌─────────────┐    ┌─────────────┐                    │
│  │  Acceptance │ → │  Onboarding │ → │  Disbursement│                    │
│  │  (e-Sign)   │    │  (Account)  │    │  (Funds)    │                    │
│  └─────────────┘    └─────────────┘    └─────────────┘                    │
│                                                                             │
└─────────────────────────────────────────────────────────────────────────────┘

# ===================================================================
# MODULE 1, LESSON 6: DIGITAL LENDING AND CREDIT
# ===================================================================

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from datetime import datetime, timedelta
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import roc_auc_score, confusion_matrix
import warnings
warnings.filterwarnings('ignore')

print("="*70)
print("DIGITAL LENDING AND CREDIT IN THE DIGITAL AGE")
print("="*70)

# ----------------------------------------------------------------
# PART A: LOAN PORTFOLIO ANALYSIS
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART A: Loan Portfolio Analysis")
print("-"*60)

# Generate synthetic loan data
np.random.seed(42)
n_loans = 10000

loan_data = pd.DataFrame({
    'loan_id': range(1, n_loans+1),
    'amount': np.random.gamma(4, 50, n_loans).clip(100, 50000),
    'term': np.random.choice([12, 24, 36, 48, 60, 72], n_loans),
    'interest_rate': np.random.uniform(0.05, 0.20, n_loans),
    'credit_score': np.random.normal(700, 50, n_loans).clip(550, 850).astype(int),
    'dti': np.random.beta(2, 5, n_loans) * 60,
    'income': np.random.gamma(5, 20, n_loans) + 20,
    'employment_years': np.random.gamma(3, 5, n_loans).clip(0, 30).astype(int),
    'home_owner': np.random.binomial(1, 0.65, n_loans),
    'purpose': np.random.choice(['debt_consolidation', 'home_improvement', 'major_purchase', 
                                 'medical', 'auto', 'other'], n_loans),
    'origination_date': [datetime(2024, 1, 1) + timedelta(days=np.random.randint(0, 365)) for _ in range(n_loans)],
    'status': np.random.choice(['Current', 'Delinquent', 'Defaulted', 'Paid Off'], n_loans, 
                              p=[0.70, 0.10, 0.05, 0.15])
})

# Generate default based on features
log_odds = (-4.5 + 0.04 * loan_data['dti'] - 0.005 * loan_data['credit_score'] 
            + 0.01 * (loan_data['amount']/1000) - 0.01 * loan_data['income']/10)
prob_default = 1 / (1 + np.exp(-log_odds))
loan_data['default'] = np.random.binomial(1, prob_default)

print(f"Generated {len(loan_data)} loans.")
print(loan_data.head())

# Portfolio summary
portfolio_summary = {
    'Total Loans': len(loan_data),
    'Total Portfolio Value': f"${loan_data['amount'].sum():,.2f}",
    'Average Loan Amount': f"${loan_data['amount'].mean():.2f}",
    'Average Interest Rate': f"{loan_data['interest_rate'].mean():.2%}",
    'Default Rate': f"{loan_data['default'].mean():.2%}",
    'Delinquent Rate': f"{(loan_data['status'] == 'Delinquent').mean():.2%}",
    'Paid Off Rate': f"{(loan_data['status'] == 'Paid Off').mean():.2%}"
}

print("\nPortfolio Summary:")
for key, value in portfolio_summary.items():
    print(f"  {key}: {value}")

# ----------------------------------------------------------------
# PART B: LOAN PERFORMANCE BY SEGMENT
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART B: Loan Performance by Segment")
print("-"*60)

# Performance by credit score segment
credit_score_bins = [550, 600, 650, 700, 750, 800, 850]
credit_score_labels = ['550-599', '600-649', '650-699', '700-749', '750-799', '800-850']
loan_data['credit_score_segment'] = pd.cut(loan_data['credit_score'], bins=credit_score_bins, labels=credit_score_labels, right=False)

segment_performance = loan_data.groupby('credit_score_segment').agg({
    'default': 'mean',
    'amount': 'mean',
    'interest_rate': 'mean'
}).round(4)

print("Loan Performance by Credit Score Segment:")
print(segment_performance)

# ----------------------------------------------------------------
# PART C: ALTERNATIVE CREDIT SCORING DEMONSTRATION
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART C: Alternative Credit Scoring Demonstration")
print("-"*60)

# Simulate alternative data features
loan_data['utility_payment_score'] = np.random.beta(2, 5, len(loan_data)) * 100
loan_data['rental_history'] = np.random.beta(3, 4, len(loan_data)) * 100
loan_data['mobile_usage'] = np.random.gamma(2, 20, len(loan_data)).clip(10, 200)
loan_data['spending_pattern'] = np.random.uniform(0.2, 0.8, len(loan_data))

# Feature engineering
features = ['credit_score', 'dti', 'income', 'employment_years', 'home_owner',
            'utility_payment_score', 'rental_history', 'mobile_usage', 'spending_pattern']
X = loan_data[features]
y = loan_data['default']

# Train-test split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)

# Train model with alternative data
model = RandomForestClassifier(n_estimators=100, max_depth=8, random_state=42)
model.fit(X_train, y_train)

# Evaluate
y_pred = model.predict_proba(X_test)[:, 1]
auc = roc_auc_score(y_test, y_pred)
print(f"Model AUC (with alternative data): {auc:.4f}")

# Feature importance
importance_df = pd.DataFrame({
    'Feature': features,
    'Importance': model.feature_importances_
}).sort_values('Importance', ascending=False)

print("\nFeature Importance:")
print(importance_df.to_string(index=False))

# ----------------------------------------------------------------
# PART D: DIGITAL ORIGINATION METRICS
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART D: Digital Origination Metrics")
print("-"*60)

# Simulate origination funnel
applications = 10000
approvals = int(applications * 0.75)
acceptances = int(approvals * 0.60)
disbursements = int(acceptances * 0.95)

origination_funnel = {
    'Stage': ['Applications', 'Approvals', 'Acceptances', 'Disbursements'],
    'Count': [applications, approvals, acceptances, disbursements],
    'Conversion Rate': [
        '100%',
        f'{approvals/applications*100:.1f}%',
        f'{acceptances/approvals*100:.1f}%',
        f'{disbursements/acceptances*100:.1f}%'
    ]
}

funnel_df = pd.DataFrame(origination_funnel)
print("Digital Origination Funnel:")
print(funnel_df.to_string(index=False))

# Visualise funnel
fig, ax = plt.subplots(figsize=(10, 6))
bars = ax.barh(funnel_df['Stage'], funnel_df['Count'], color='steelblue', alpha=0.7)
ax.set_xlabel('Count')
ax.set_title('Digital Origination Funnel')
for bar, count in zip(bars, funnel_df['Count']):
    ax.text(bar.get_width() + 200, bar.get_y() + bar.get_height()/2, 
            str(count), ha='left', va='center', fontweight='bold')
ax.grid(True, alpha=0.3)
plt.tight_layout()
plt.savefig('origination_funnel.png', dpi=300, bbox_inches='tight')
plt.show()
print("Origination funnel visualisation saved as 'origination_funnel.png'")

# ----------------------------------------------------------------
# PART E: RISK-BASED PRICING SIMULATION
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART E: Risk-Based Pricing Simulation")
print("-"*60)

# Simulate risk-based pricing
risk_bins = [0, 0.05, 0.10, 0.15, 0.20, 0.30, 1.0]
risk_labels = ['Very Low', 'Low', 'Medium-Low', 'Medium', 'Medium-High', 'High']
loan_data['risk_segment'] = pd.cut(loan_data['default'], bins=risk_bins, labels=risk_labels, right=False)

# Pricing by risk segment
pricing_rates = {
    'Very Low': 0.045,
    'Low': 0.06,
    'Medium-Low': 0.08,
    'Medium': 0.10,
    'Medium-High': 0.14,
    'High': 0.20
}

loan_data['risk_rate'] = loan_data['risk_segment'].map(pricing_rates)

pricing_summary = loan_data.groupby('risk_segment').agg({
    'default': 'mean',
    'amount': 'mean',
    'risk_rate': 'first'
}).round(4)

print("Risk-Based Pricing Summary:")
print(pricing_summary)

# ----------------------------------------------------------------
# PART F: DIGITAL LENDING STRATEGY
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART F: Digital Lending Strategy")
print("-"*60)

strategy = {
    "1. Customer Acquisition": {
        "Tactics": [
            "Digital marketing for specific segments.",
            "Referral programs.",
            "Partnerships with fintechs.",
            "Embedded lending in partner platforms."
        ],
        "Priority": "High",
        "Timeline": "Now"
    },
    "2. Origination": {
        "Tactics": [
            "Mobile-first application process.",
            "Instant decisioning with AI.",
            "E-signature and digital onboarding.",
            "API integration for data verification."
        ],
        "Priority": "High",
        "Timeline": "Now"
    },
    "3. Credit Scoring": {
        "Tactics": [
            "Implement alternative credit scoring.",
            "Use ML for risk assessment.",
            "Regular model validation and monitoring.",
            "Fairness and bias testing."
        ],
        "Priority": "High",
        "Timeline": "6 months"
    },
    "4. Portfolio Management": {
        "Tactics": [
            "Real-time portfolio monitoring.",
            "Predictive analytics for early warning.",
            "Automated collections.",
            "Stress testing and scenario analysis."
        ],
        "Priority": "Medium",
        "Timeline": "12 months"
    },
    "5. Innovation": {
        "Tactics": [
            "Explore BNPL integration.",
            "Embedded lending in e-commerce.",
            "Use generative AI for customer engagement.",
            "Explore blockchain for credit."
        ],
        "Priority": "Medium",
        "Timeline": "18 months"
    }
}

print("Digital Lending Strategy:")
for strategy_item, details in strategy.items():
    print(f"\n{strategy_item}:")
    print("  Tactics:")
    for tactic in details['Tactics']:
        print(f"    • {tactic}")
    print(f"  Priority: {details['Priority']}")
    print(f"  Timeline: {details['Timeline']}")

# ----------------------------------------------------------------
# PART G: SUMMARY AND RECOMMENDATIONS
# ----------------------------------------------------------------

print("\n" + "="*70)
print("PART G: Summary and Recommendations")
print("="*70)

print("""
Digital Lending – Key Takeaways:

1. Digital lending encompasses personal loans, SME lending, mortgages, BNPL, and more.
2. Alternative credit scoring uses non-traditional data (utilities, rent, transactions).
3. AI-powered credit scoring improves accuracy and inclusivity.
4. Digital origination is mobile-first, instant, and paperless.
5. Risk management includes credit risk, fraud risk, model risk, and regulatory risk.
6. Regulatory compliance is essential (ECOA, GDPR, IFRS 9, Basel III).
7. Digital lending strategy must balance growth, risk, and compliance.

Recommendations:
  - Implement alternative credit scoring.
  - Build a mobile-first digital origination platform.
  - Invest in AI for risk assessment and decisioning.
  - Ensure regulatory compliance (fair lending, data privacy).
  - Develop risk-based pricing models.
  - Partner with fintechs and data providers.
""")

print("="*70)
print("END OF LESSON 6 – MODULE 1")
print("="*70)