create_samples.py

# Create sample CSV files for testing the fraud detection modules
import pandas as pd
import numpy as np
from datetime import datetime, timedelta

# 1. Sample Transaction Data
def create_transaction_sample():
    np.random.seed(42)
    data = {
        'transaction_id': [f'TXN_{i:06d}' for i in range(1, 1001)],
        'customer_id': [f'CUST_{i:04d}' for i in np.random.randint(1, 501, 1000)],
        'amount': np.random.exponential(100, 1000),
        'merchant_category': np.random.choice(['GROCERY', 'GAS', 'RESTAURANT', 'ONLINE', 'HIGH_RISK', 'ATM'], 1000),
        'timestamp': pd.date_range(start='2024-01-01', periods=1000, freq='H'),
        'currency': 'USD'
    }
    
    # Add some fraudulent patterns
    fraud_indices = np.random.choice(1000, 50, replace=False)
    data['amount'][fraud_indices] = np.random.uniform(15000, 50000, 50)  # Large amounts
    
    # Add negative amounts (errors)
    error_indices = np.random.choice(1000, 10, replace=False)
    data['amount'][error_indices] = -np.random.uniform(100, 1000, 10)
    
    df = pd.DataFrame(data)
    df.to_csv('sample_transactions.csv', index=False)
    return df

# 2. Sample KYC Data
def create_kyc_sample():
    np.random.seed(42)
    names = ['John Smith', 'Jane Doe', 'Michael Johnson', 'Sarah Wilson', 'David Brown', 
             'JOHN SMITH', 'Test123', 'A B', 'Maria Garcia', 'James Miller']
    
    data = {
        'customer_id': [f'CUST_{i:04d}' for i in range(1, 101)],
        'name': np.random.choice(names, 100),
        'email': [f'user{i}@email.com' for i in range(1, 101)],
        'phone': [f'+1555{i:07d}' for i in range(1, 101)],
        'dob': pd.date_range(start='1950-01-01', end='2005-12-31', periods=100).strftime('%Y-%m-%d'),
        'address': [f'{i} Main St, City, State' for i in range(1, 101)]
    }
    
    # Add duplicates
    data['email'][5:10] = 'duplicate@email.com'
    data['phone'][15:18] = '+15551234567'
    
    # Add invalid DOBs
    data['dob'][20] = '2030-01-01'  # Future date
    data['dob'][21] = 'invalid-date'  # Invalid format
    data['dob'][22] = '1900-01-01'  # Very old
    
    df = pd.DataFrame(data)
    df.to_csv('sample_kyc.csv', index=False)
    return df

# 3. Sample Sanctions List
def create_sanctions_sample():
    sanctions_data = {
        'name': [
            'John Doe', 'Jane Smith', 'Vladimir Putin', 'Kim Jong Un',
            'Alexander Petrov', 'Maria Gonzalez', 'Ahmed Hassan',
            'Natasha Volkov', 'Carlos Rodriguez', 'Wei Zhang'
        ],
        'list_type': [
            'OFAC', 'EU', 'UN', 'UN', 'OFAC', 'EU', 'OFAC', 'EU', 'OFAC', 'UN'
        ],
        'date_added': pd.date_range(start='2020-01-01', periods=10, freq='30D')
    }
    
    df = pd.DataFrame(sanctions_data)
    df.to_csv('sample_sanctions.csv', index=False)
    return df

# 4. Sample Customer List for Sanctions Screening
def create_customer_sample():
    np.random.seed(42)
    customer_names = [
        'John Doe', 'Alice Johnson', 'Bob Wilson', 'Jane Smith',
        'Michael Brown', 'Sarah Davis', 'Vladimir Putin', 'Emily Chen',
        'David Miller', 'Lisa Anderson', 'Ahmed Hassan', 'Maria Lopez'
    ]
    
    data = {
        'customer_id': [f'CUST_{i:04d}' for i in range(1, 101)],
        'name': np.random.choice(customer_names, 100),
        'dob': pd.date_range(start='1950-01-01', end='2000-12-31', periods=100).strftime('%Y-%m-%d'),
        'country': np.random.choice(['USA', 'UK', 'Canada', 'Germany', 'France'], 100)
    }
    
    df = pd.DataFrame(data)
    df.to_csv('sample_customers.csv', index=False)
    return df

# 5. Sample Credit Risk Data
def create_credit_sample():
    np.random.seed(42)
    data = {
        'customer_id': [f'CUST_{i:04d}' for i in range(1, 201)],
        'credit_score': np.random.normal(700, 100, 200).astype(int),
        'utilization_rate': np.random.beta(2, 5, 200),
        'debt_to_income': np.random.beta(2, 3, 200),
        'income': np.random.lognormal(10.5, 0.5, 200).astype(int),
        'recent_defaults': np.random.poisson(0.1, 200),
        'credit_history_length': np.random.randint(1, 30, 200)
    }
    
    # Create some high-risk profiles
    high_risk_indices = np.random.choice(200, 30, replace=False)
    data['credit_score'][high_risk_indices] = np.random.randint(300, 600, 30)
    data['utilization_rate'][high_risk_indices] = np.random.uniform(0.8, 1.0, 30)
    data['debt_to_income'][high_risk_indices] = np.random.uniform(0.4, 0.8, 30)
    
    df = pd.DataFrame(data)
    df.to_csv('sample_credit.csv', index=False)
    return df

# Create all sample files
if __name__ == "__main__":
    print("Creating sample CSV files...")
    create_transaction_sample()
    create_kyc_sample() 
    create_sanctions_sample()
    create_customer_sample()
    create_credit_sample()
    print("Sample files created successfully!")
    print("\nFiles created:")
    print("- sample_transactions.csv")
    print("- sample_kyc.csv") 
    print("- sample_sanctions.csv")
    print("- sample_customers.csv")
    print("- sample_credit.csv")