@forestadmin/datasource-demo-fintech

Demo data source for Forest Admin, themed around a fintech / compliance back office (customers, cards, KYC, AML alerts, chargebacks…).

On boot it generates a throwaway SQLite database with faker, seeds it via Knex, and exposes it through @forestadmin/datasource-sql. The schema, column types and primary keys are introspected from the freshly-created tables. It is used by forest create demo to bootstrap a project with realistic sample data.

import { createDemoFintechDataSource } from '@forestadmin/datasource-demo-fintech';

agent.addDataSource(createDemoFintechDataSource());

How the data is built

A fresh database is generated on every boot (no fixed seed → the dataset is re-rolled each time), under the OS temp directory. The seed lives in src/seed:

• seed/index.ts — orchestrates the builders in dependency order (customers first, then everything that references them) and owns the Knex connection.
• one file per domain (customers, cards, kyc, aml, refunds, chargebacks) — each creates its table(s) and inserts rows.
• pools.ts / utils.ts — curated reference data (kept hand-written so geography stays internally consistent) and faker-backed random helpers.

Business coherence is enforced in the builders, not by faker: sanctioned customers never clear onboarding, alerts are weighted by customer risk, a filed SAR escalates its source alert, chargebacks link back to the right card and refund request, etc.

Relations, Smart Actions, segments and computed fields are layered on top in src/customizations/*. Dates use the relative daysAgo(n) / dayOnlyAgo(n) helpers (see src/dates.ts) so the demo always shows recent data.

To rescale the dataset, edit the population sizes in the relevant src/seed/* builder (e.g. N in seed/customers.ts).