Thirty years. A clearer picture.
A national map makes local history look easy. The hard part is that thirty years of consistent data does not exist at every geographic level. The product needed to answer useful questions without presenting regional averages as station-level facts.
Decisions that
shape the system.
Start with a defensible baseline
Use annual state history as the long-range backbone and add higher-frequency series where published sources support them.
Make comparison approachable
Connect map navigation, location search, a timeline, and a comparison tray. Limit active overlays to five so the chart remains readable.
Carry provenance through the stack
Shared contracts preserve direct-local, direct-regional, fallback, and unsupported coverage classes from ingestion to the interface.
Keep the architecture adaptable
Separate data acquisition, shared schemas, serving logic, and presentation. The API supports a compiled-data path and an optional PostgreSQL backend.
- Source snapshots
- Validated facts
- Typed API
- Map & timeline
- Location comparison
Inside the implementation
The original request centered on layperson-friendly comparisons. The implementation uses long-range line charts for trends and grouped comparisons for narrower windows. Viewport-driven map requests and capped responses keep interaction bounded, while downsampling supports long weekly series. A county or ZIP search can resolve to a supported parent region with that relationship explicitly displayed.
What the work
demonstrates.
The workspace contains a responsive map-and-chart application, typed API, shared validation schemas, source refresh adapters, compiled datasets, and operational runbooks. EIA and BLS price series have distinct roles; FHWA tax tables remain contextual data.
Why it matters to a client
Useful for businesses that need to turn fragmented datasets into a decision tool, with understandable comparisons and traceable methodology.