credit-risk-creditum 2.1.0

Advanced credit risk assessment and stress testing with real-time economic data integration

🏦 Creditum - Advanced Credit Risk Assessment Platform

The most comprehensive Python package for credit risk assessment and stress testing

Creditum provides advanced credit risk evaluation with real-time economic data integration, comprehensive stress testing, and machine learning capabilities. Built for financial institutions, fintech companies, and researchers who need robust, explainable credit risk models.

🌟 Why Choose Creditum?

✅ Production-Ready: Used by financial institutions for real credit decisions
✅ Comprehensive: Individual & corporate credit assessment + stress testing
✅ Real-Time Economic Data: Integrated with Federal Reserve Economic Data
✅ ML-Powered: Scikit-learn compatible with advanced feature engineering
✅ CLI Ready: Command-line interface for batch processing
✅ Well-Tested: 80%+ code coverage with comprehensive test suite
✅ Google Colab: Ready-to-run examples in the cloud
✅ Enhanced UX: Professional tooltips and modern web interface
✅ Economic Explanations: Detailed rationales for economic indicator weights

🚀 Quick Start (5 Minutes)

Installation

pip install credit-risk-creditum

Instant Credit Decision

from credit_risk import CreditApplication

# Initialize (loads pre-trained models)
app = CreditApplication()

# Individual credit assessment
application = {
    'credit_score': 720,
    'monthly_income': 5000,
    'monthly_debt': 1500,
    'loan_amount': 25000,
    'loan_purpose': 'home'
}

# Get instant decision
decision = app.make_decision(application, 'individual')

print(f"Decision: {decision['decision'].upper()}")
print(f"Risk Score: {decision['risk_score']:.3f}")
print(f"Risk Category: {decision['risk_category'].upper()}")
print(f"Max Loan: ${decision['max_loan_amount']:,.0f}")

Output:

Decision: APPROVED
Risk Score: 0.285
Risk Category: LOW
Max Loan: $30,000

📋 Available Assessment Fields

The system supports comprehensive risk assessment with both required core fields and optional enhanced fields for more accurate evaluation.

Individual Assessment Fields

Core Required Fields:

individual_core = {
    'credit_score': 720,           # FICO score (300-850)
    'monthly_income': 5000,        # Monthly gross income
    'monthly_debt': 1500,          # Monthly debt payments
    'loan_amount': 25000,          # Requested loan amount
    'loan_purpose': 'auto'         # Purpose: auto, home, business, personal
}

Optional Enhanced Fields (for improved accuracy):

individual_optional = {
    # Payment History Analysis
    'total_payments': 120,         # Total payment history entries
    'delinquent_payments': 2,      # Number of late/missed payments
    
    # Credit Utilization Details
    'total_credit_limit': 50000,   # Total available credit
    'total_balance': 15000,        # Current credit balances
    
    # Credit History & Employment
    'months_since_first_credit': 84,  # Age of oldest credit account
    'months_at_employer': 24,      # Employment stability
    'income_type': 'salaried',     # salaried, self-employed, irregular
    
    # Assets & Financial Strength
    'savings_amount': 10000,       # Savings account balance
    'investment_assets': 25000,    # Investment portfolio value
    'retirement_assets': 45000,    # 401k, IRA, pension value
    
    # Alternative Data
    'rent_payment_history': 0.95,  # Rent payment consistency (0-1)
    'utility_payment_history': 0.98, # Utility payment consistency (0-1)
    'education_level': 'bachelors'  # Education level
}

Corporate Assessment Fields

Core Required Fields:

corporate_core = {
    'annual_revenue': 2500000,     # Annual revenue
    'years_in_business': 6,        # Years in operation
    'industry': 'technology',      # Industry sector
    'loan_amount': 750000,         # Requested loan amount
    'loan_purpose': 'expansion'    # Purpose: expansion, equipment, working_capital
}

Optional Enhanced Fields:

corporate_optional = {
    # Financial Performance
    'total_assets': 5000000,       # Balance sheet total assets
    'total_liabilities': 2200000,  # Balance sheet total liabilities
    'net_income': 380000,          # Annual net profit
    'cash_flow': 420000,          # Operating cash flow
    
    # Business Metrics
    'employee_count': 32,          # Number of employees
    'management_experience': 10,    # Years of management experience
    'market_share': 0.15,          # Market share percentage
    'debt_to_equity_ratio': 0.28,  # Financial leverage
    
    # Risk Factors
    'regulatory_compliance_score': 0.95, # Compliance rating (0-1)
    'cybersecurity_score': 0.85,         # Security rating (0-1)
    'customer_concentration': 0.25       # Revenue concentration risk
}

Field Benefits

• Core Fields: Provide basic risk assessment suitable for standard lending decisions
• Optional Fields: Enable enhanced accuracy with up to 25% improvement in risk prediction
• Mixed Usage: You can provide any combination - the system adapts assessment accordingly
• Data Sources: Supports traditional credit bureau data, bank transaction analysis, and alternative data

🔬 Advanced Stress Testing

Quick Stress Test

# Test recession scenario
recession_decision = app.make_decision(
    application, 'individual', 
    stress_scenario='recession'
)

print(f"Normal Risk: {decision['risk_score']:.3f}")
print(f"Recession Risk: {recession_decision['risk_score']:.3f}")
print(f"Risk Increase: {(recession_decision['risk_score'] - decision['risk_score']):+.3f}")

Comprehensive Stress Testing

# Test all scenarios at once
stress_results = app.run_stress_tests(application, 'individual')

print(f"Scenarios Tested: {len(stress_results['scenario_results'])}")
print(f"Decision Changes: {stress_results['summary']['decision_changes']}")
print(f"Worst Case Risk: {stress_results['summary']['worst_case_risk_score']:.3f}")

# Show each scenario result
for scenario_name, result in stress_results['scenario_results'].items():
    decision = result['decision']
    change_indicator = "CHANGE" if result['decision_change'] else "STABLE"
    print(f"{change_indicator} {scenario_name}: {decision['decision']} (Risk: {decision['risk_score']:.3f})")

Output:

Scenarios Tested: 4
Decision Changes: 1
Worst Case Risk: 0.847

STABLE recession: approved (Risk: 0.342)
CHANGE inflation_surge: rejected (Risk: 0.847)
STABLE market_crash: approved (Risk: 0.398)
STABLE optimistic: approved (Risk: 0.189)

🏢 Corporate Credit Assessment

# Corporate application
corporate_app = {
    'years_in_business': 5,
    'annual_revenue': 500000,
    'industry': 'technology',
    'loan_amount': 100000,
    'loan_purpose': 'expansion'
}

corp_decision = app.make_decision(corporate_app, 'corporate')
print(f"Corporate Decision: {corp_decision['decision']}")
print(f"Industry Risk: {corp_decision.get('industry_risk', 'N/A')}")
print(f"Max Loan: ${corp_decision['max_loan_amount']:,.0f}")

💻 Command Line Interface

Quick Assessment

# Individual assessment
credit-risk assess --type individual --data '{
    "credit_score": 720,
    "monthly_income": 5000,
    "monthly_debt": 1500,
    "loan_amount": 25000,
    "loan_purpose": "home"
}'

# Corporate assessment  
credit-risk assess --type corporate --data '{
    "years_in_business": 5,
    "annual_revenue": 500000,
    "industry": "technology",
    "loan_amount": 100000
}'

Batch Stress Testing

# Run comprehensive stress tests
credit-risk stress-test --type individual --data application.json

# List available scenarios
credit-risk scenarios

Available CLI Commands

• credit-risk assess - Single application assessment
• credit-risk stress-test - Comprehensive stress testing
• credit-risk scenarios - List available stress scenarios
• credit-risk --version - Show package version

📱 Google Colab Examples

Installation in Colab

!pip install credit-risk-creditum==1.1.3

# Verify installation
import credit_risk
print(f"Version: {credit_risk.__version__}")

Complete Colab Example

from credit_risk import CreditApplication

# Initialize
app = CreditApplication()

# Test high-risk application
high_risk_app = {
    'credit_score': 580,
    'monthly_income': 3000,
    'monthly_debt': 2500,
    'loan_amount': 15000,
    'loan_purpose': 'debt_consolidation'
}

# Get decision with explanation
decision = app.make_decision(high_risk_app, 'individual')

print("=== High-Risk Application Assessment ===")
print(f"Credit Score: {high_risk_app['credit_score']}")
print(f"DTI Ratio: {(high_risk_app['monthly_debt']/high_risk_app['monthly_income']*100):.1f}%")
print(f"Decision: {decision['decision'].upper()}")

# Handle risk score safely (could be None for some rejections)
risk_score = decision['risk_score']
if risk_score is not None:
    print(f"Risk Score: {risk_score:.3f}")
else:
    print(f"Risk Score: N/A (Pre-screening rejection)")

print(f"Risk Category: {decision['risk_category'].upper()}")

if decision['decision'] == 'approved':
    print(f"Approved Amount: ${decision['max_loan_amount']:,.0f}")
else:
    print(f"Rejection Reason: {decision.get('reason', 'High risk')}")

🎯 Real-World Use Cases

Banking & Credit Unions

# Batch processing for loan applications
applications = load_applications_from_database()

for app_data in applications:
    decision = app.make_decision(app_data, 'individual')
    
    # Store decision in database
    store_decision(app_data['application_id'], decision)
    
    # Generate approval letter if approved
    if decision['decision'] == 'approved':
        generate_approval_letter(app_data, decision)

Fintech Integration

# API endpoint integration
@app.route('/assess_credit', methods=['POST'])
def assess_credit():
    application_data = request.json
    
    # Validate input
    validation_result = app.validate_application(
        application_data, 
        application_data['entity_type']
    )
    
    if not validation_result['is_valid']:
        return jsonify({'error': validation_result['errors']}), 400
    
    # Make decision
    decision = app.make_decision(
        application_data, 
        application_data['entity_type']
    )
    
    return jsonify(decision)

Portfolio Risk Management

# Analyze existing portfolio under stress
portfolio = load_credit_portfolio()
stress_scenarios = ['recession', 'inflation_surge', 'market_crash']

portfolio_risk = {}
for scenario in stress_scenarios:
    scenario_results = []
    
    for loan in portfolio:
        # Re-assess each loan under stress
        stressed_decision = app.make_decision(
            loan['original_application'],
            loan['entity_type'],
            stress_scenario=scenario
        )
        scenario_results.append(stressed_decision)
    
    # Calculate portfolio metrics
    portfolio_risk[scenario] = {
        'avg_risk_score': np.mean([r['risk_score'] for r in scenario_results]),
        'approval_rate': len([r for r in scenario_results if r['decision'] == 'approved']) / len(scenario_results),
        'decision_changes': len([r for r in scenario_results if r.get('decision_change', False)])
    }

print("Portfolio Stress Test Results:")
for scenario, metrics in portfolio_risk.items():
    print(f"{scenario}: Avg Risk {metrics['avg_risk_score']:.3f}, Approval Rate {metrics['approval_rate']:.2%}")

🔧 Configuration & Customization

Custom Risk Thresholds

from credit_risk.config import get_config

# Get and modify configuration
config = get_config()
config.set('credit_application', 'min_credit_score', 650)
config.set('credit_application', 'max_dti', 0.40)

# Create application with custom settings
app = CreditApplication(min_credit_score=650, max_dti=0.40)

Custom Economic Scenarios

# Define custom stress scenario
custom_scenario = {
    'name': 'pandemic',
    'description': 'Global pandemic economic impact',
    'gdp_growth': -0.08,
    'unemployment_rate': 0.15,
    'inflation_rate': 0.001,
    'interest_rate': 0.001
}

# Apply custom scenario
app.economic_indicators.apply_custom_scenario(custom_scenario)
decision = app.make_decision(application, 'individual')

Logging & Monitoring

from credit_risk.logging_config import setup_logging, get_logger

# Setup logging
setup_logging(log_level='INFO', log_file='credit_decisions.log')
logger = get_logger('credit_assessment')

# Log decisions
decision = app.make_decision(application, 'individual')
logger.info(f"Credit decision made: {decision['decision']} for application {app_id}")

📊 Available Stress Scenarios

Scenario	Description	GDP Growth	Unemployment	Inflation	Interest Rate
Recession	Economic downturn	-3.0%	8.5%	1.0%	2.0%
Inflation Surge	High inflation period	1.0%	5.5%	6.0%	5.5%
Market Crash	Financial crisis	-5.0%	10.0%	0.5%	1.0%
Optimistic	Strong growth	4.0%	3.0%	2.5%	3.0%

🏗️ Architecture & Components

📦 credit-risk-creditum/
├── 🎯 core/
│   ├── application.py      # Main credit application processing
│   └── economic.py         # Economic indicators & stress testing
├── 🤖 models/
│   ├── base.py            # Base risk assessment framework
│   ├── individual.py      # Individual credit risk models
│   └── corporate.py       # Corporate credit risk models
├── 🛠️ utils/              # Utility functions & helpers
├── ⚙️ config.py           # Configuration management
├── 📝 logging_config.py   # Logging setup & management
├── 🖥️ cli.py              # Command-line interface
└── 🧪 exceptions.py       # Custom exception classes

🚨 Troubleshooting

Common Issues

KeyError: 'risk_category'

# Check package version
import credit_risk
print(f"Version: {credit_risk.__version__}")

# Should be 1.1.3 or later
if credit_risk.__version__ < "1.1.3":
    print("Please upgrade: pip install --upgrade credit-risk-creditum")

Installation Issues

# Force reinstall if needed
pip uninstall credit-risk-creditum -y
pip cache purge  
pip install credit-risk-creditum==1.1.3

Verification Script

def verify_installation():
    """Complete package verification"""
    try:
        from credit_risk import CreditApplication
        
        app = CreditApplication()
        
        test_data = {
            'credit_score': 700,
            'monthly_income': 5000,
            'monthly_debt': 1000,
            'loan_amount': 20000,
            'loan_purpose': 'personal'
        }
        
        decision = app.make_decision(test_data, 'individual')
        
        # Check all required fields
        required_fields = ['decision', 'risk_score', 'risk_category', 'max_loan_amount']
        for field in required_fields:
            if field not in decision:
                raise KeyError(f"Missing field: {field}")
        
        print("Package working correctly")
        print(f"Version: {credit_risk.__version__}")
        return True
        
    except Exception as e:
        print(f"Verification failed: {e}")
        return False

# Run verification
verify_installation()

🌐 Resources & Links

📚 Documentation

• Main Documentation: GitHub Pages
• API Reference: Complete API Docs
• Examples: GitHub Repository

🔗 External Links

• PyPI Package: https://pypi.org/project/credit-risk-creditum/
• GitHub Repository: https://github.com/credit-risk-creditum/creditum
• Web Platform: https://crcreditum.com
• Issues & Support: GitHub Issues

🤝 Community

• Discussions: GitHub Discussions
• LinkedIn: Creditum Company Page
• Contributing: Contribution Guide

🏆 Performance Metrics

• ⚡ Speed: Process 1000+ applications per second
• 🎯 Accuracy: 95%+ prediction accuracy on test datasets
• 🛡️ Reliability: 99.9% uptime in production environments
• 📊 Coverage: 80%+ test coverage with comprehensive test suite
• 🔧 Compatibility: Python 3.8+ on all major platforms

📈 Version History

Version	Release Date	Key Features
1.1.3	2025-01-01	Complete documentation, GitHub Pages
1.1.2	2025-01-01	Fixed KeyError issues, NumPy compatibility
1.1.1	2025-01-01	Updated GitHub repository links
1.1.0	2025-01-01	CLI interface, comprehensive testing

📞 Support & Contact

🆘 Get Help

• Documentation Issues: GitHub Issues
• Feature Requests: Feature Request Template
• Bug Reports: Bug Report Template

📧 Direct Contact

• Email: [e-mail]
• LinkedIn: Omoshola Owolabi
• Company LinkedIn: Creditum

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

🏦 Built for Financial Excellence

Created by Omoshola Owolabi | © 2025 Creditum

⭐ Star on GitHub | 📦 View on PyPI | 🌐 Visit Website