Case Study
FusionScope
Macro-risk dashboard fusing live keyless ETL (World Bank, GDACS, UNHCR) across 60 countries into an INFORM-aligned, geometric-mean 3-dimension risk score with a real historical trend store.
Executive Summary
FusionScope is a full-stack macro-stability monitor: an interactive globe over ~60 countries, fed by live keyless public data sources, scored with an INFORM-aligned methodology instead of an ad-hoc weighted sum.
Problem & Constraints
Macro risk monitoring means pulling heterogeneous global indicators into one geospatial dashboard — without paid data feeds, and without pretending mock numbers are live.
Architecture
Keyless ETL (World Bank indicators + GDACS disaster alerts + UNHCR displacement) → atomic refresh → INFORM-aligned 3-dimension scoring (hazard / vulnerability / coping capacity, geometric mean) → append-only history store → FastAPI → React globe + alerts + feed.
Methodology
- Live ETL from three keyless public sources with per-source status reporting and outage fallback
- Risk scoring aligned to the INFORM framework: three dimensions combined by geometric mean, not a flat average
- Append-only history store gives real trend lines instead of synthetic sparklines
- Alerts and Feed pages wired to the real API (previously mock-only)
- Atomic refresh so a partial source failure never serves a half-updated dataset
Results & Metrics
| Metric | Result |
|---|---|
| Live sources | World Bank + GDACS + UNHCR (keyless) |
| Coverage | ~60 countries |
| Scoring | INFORM 3-dim geometric mean |
| Tests | 38 backend (pytest) + 20 frontend |
Tech Stack
React, TypeScript, Vite, shadcn/ui, Python FastAPI
Future Work
Deployment (pending), ML-based anomaly detection, historical event correlation.