Tekhnika/fraud-model-starter-pack-free

Fraud Model Starter Pack

Free sample for fraud monitoring dashboards, suspicious activity analysis, and finance-oriented analytics workflows.

What is included

• card_transactions.csv: 4398 rows, 10 columns
• cardholders.csv: 439 rows, 9 columns
• cards.csv: 593 rows, 9 columns
• daily_fraud_metrics.csv: 733 rows, 10 columns
• fraud_cases.csv: 967 rows, 10 columns
• merchant_risk_profiles.csv: 366 rows, 9 columns

Why this dataset is useful

• Useful for a first fraud dashboard or suspicious activity notebook.
• Works well for SQL, notebooks, and BI prototyping.
• Provides a reduced but representative sample of the core workflow in the full starter pack.

Starter use cases

• Fraud baseline using linked workflow and event data.
• Fraud monitoring dashboard for suspicious activity patterns.

Schema overview

card_transactions.csv

• Rows: 4398
• Columns: transaction_id, cardholder_id, card_id, merchant_id, transaction_date, merchant_category, transaction_amount, merchant_country, entry_mode, transaction_currency

cardholders.csv

• Rows: 439
• Columns: cardholder_id, credit_score_band, customer_name, home_region, customer_segment, tenure_months, travel_frequency, historical_chargeback_flag, risk_band

cards.csv

• Rows: 593
• Columns: card_id, cardholder_id, issue_date, status, card_product, network, credit_limit, card_present_usage_ratio, digital_wallet_flag

daily_fraud_metrics.csv

• Rows: 733
• Columns: metric_id, merchant_id, metric_date, transactions_total, approved_amount, decline_rate, fraud_alerts_total, confirmed_fraud_total, chargeback_rate, loss_amount

fraud_cases.csv

• Rows: 967
• Columns: fraud_case_id, transaction_id, cardholder_id, card_id, alert_date, fraud_type, investigation_status, chargeback_amount, loss_amount, model_score

merchant_risk_profiles.csv

• Rows: 366
• Columns: merchant_id, merchant_category, merchant_country, merchant_name, acquirer_region, merchant_size, card_not_present_share, chargeback_ratio, merchant_risk_band

Free vs full version

• Free Kaggle sample: reduced rows, reduced columns, starter notebook, and enough linked finance fraud tables to validate the core workflow.
• Full version: full row volume, richer feature coverage, and extra starter assets for dashboards, SQL, and fraud-analysis work.

Upgrade to full version

• Full version: https://tekhnikalab.gumroad.com/l/fraud-model-starter-pack
• Upgrade if you need the full linked schema plus starter assets that get you to a fraud dashboard, SQL project, or risk baseline faster.

Notes

• Contains generated data only and no real personal data.
• Designed as a lightweight free sample for evaluation and discovery.