~/cases/open-earthLive
Open Earth
How do you combine heterogeneous scientific datasets and live observations without erasing provenance, uncertainty or scientific meaning?
→Context
Open Earth brings earthquakes, volcanoes and tectonic context into one inspectable map.
→Data model
Reference datasets, live observations and derived spatial relationships use separate paths and freshness semantics.
→Architecture
Python preparation produces static snapshots and manifests; browser adapters render MapLibre layers, search and an inspector.
→Trade-offs
Static delivery avoids a runtime backend, but scheduled updates and explicit source metadata become part of the product.
- Scientific sources are normalized into browser-safe snapshots and a manifest.
- The browser combines snapshots with live USGS observations.
- The interface preserves provenance, freshness and distinctions between reference, observed and derived information.
→Decisions & trade-offs
| Decision | Why | Trade-off |
|---|---|---|
| Keep provenance in the normalized data | Users need to know where a feature came from and how fresh it is. | Data and UI models are more explicit. |
| Make map state URL-first | Views and selected features remain shareable and recoverable. | State serialization must stay stable. |
→Result
- A static, source-aware Earth-science explorer is deployed at open-earth.info.
- The repository includes JavaScript and Python tests for state, data and regional adapters.
→Evidence
- Live/reference sources
- USGS · GVP
- Map rendering
- MapLibre
- Deployment model
- Static
Earthquakes and volcano datasets
Native browser ES modules
No account, backend or API key
→What I'd change today
Source freshness should be more visible without overwhelming the map.