Hospigen — Canadian healthcare synthetic data & analytics

This repository contains a small, opinionated toolset for generating synthetic Canadian healthcare data (via Synthea), publishing it into a Google Cloud FHIR Store, and turning those clinical events into analytics-ready BigQuery tables for dashboards and ML.

Two primary entry points:

• Analytics (generation): see analytics/generation/README.md — the generation pipeline that runs Synthea, uploads per-patient NDJSON to GCS, loads into a BigQuery staging table, and materializes analytic tables (e.g., patients) for dashboards and machine learning.
• Services (FHIR bridge): see services/bridge/bridge_README.md — a FastAPI bridge that listens to FHIR Store change notifications, fetches full FHIR resources, wraps them as business events, and publishes to Pub/Sub.

Scope

• In /services we run Synthea and operate the Bridge to populate a Google Cloud FHIR Store and emit downstream event streams.
• In /analytics/generation we build the analytics pipeline that ingests GCS / FHIR events into BigQuery for reporting, dashboards, and ML experiments.

Next steps / roadmap

• Create a more "living hospital system" where events naturally occur as time passes (continuous time simulation).
• Modify Synthea modules and orchestration so protocol-driven scenarios can be played (e.g., outbreak, surge, clinical pathways).
• Add real-time (RT) pipelines and streaming metrics so dashboards can surface signals like "Average real-time heart rate in the hospital" (haha).
• Implement remote patient monitoring (RPM) scenarios that combine RT or batch ingestion with ML / Generative AI analytics on top for anomaly detection, risk scoring, or decision support.

Context

• This work is oriented to Canadian healthcare settings and uses Canadian defaults and assumptions where applicable.

See also

• services/bridge/bridge_README.md
• analytics/generation/README.md