How anant2025grad/finguard works
FinGuard is a single-file synthetic financial transaction data generator, not a transaction-processing service. Its principal workflow creates account and transaction data, injects labeled fraud scenarios, engineers model-ready features, validates the resulting dataset, and saves it. The excerpt documents these stages and the script entry point; omitted portions limit certainty about some internal call details and the output path.
Main entry points: data_generator.py Generation entry point
Components 9
Other components
- Script user
- Generation entry point
src/data_generator.py - Account generation
src/data_generator.py - Transaction generation
src/data_generator.py - Fraud injection
src/data_generator.py - Feature engineering
src/data_generator.py - Dataset validation
src/data_generator.py - Dataset saving
src/data_generator.py - Generated dataset
Connections 9
- Script user → Generation entry point: runs
- Generation entry point → Account generation: generates
- Generation entry point → Transaction generation: generates
- Account generation → Transaction generation: supplies accounts
- Generation entry point → Fraud injection: injects fraud
- Generation entry point → Feature engineering: engineers features
- Generation entry point → Dataset validation: validates data
- Generation entry point → Dataset saving: saves data
- Dataset saving → Generated dataset: writes