Skip to content

Latest commit

 

History

History
27 lines (19 loc) · 838 Bytes

File metadata and controls

27 lines (19 loc) · 838 Bytes

Data Provenance

All datasets used by this project are generated synthetically by the code in this repository.

The synthetic data is designed to represent plausible financial-services behaviour for demonstration purposes only.

No external, confidential, employer, customer, member, operational, financial, or proprietary datasets are used.

Synthetic data layers

The project generates synthetic data for:

  • Customer profiles.
  • Monthly relationship activity.
  • Digital engagement.
  • Product relationship depth.
  • Credit relationship signals.
  • Service interaction signals.
  • Process friction signals.
  • Voluntary attrition labels.
  • Intervention assignments.
  • Treatment logs.
  • Post-intervention outcomes.

Reproducibility

Synthetic data generation uses deterministic random seeds so that results can be reproduced locally.