Selected Project · Private Prototype
GovDataHub
A production-oriented data-as-a-service layer that makes public U.S. Census data easier to discover, query, and reuse.
The implementation repository is private. This page documents the service design and public-data reuse model without exposing private source code.
The Problem
Government data is public, but working with it often means jumping between dataset portals, learning a new query shape for each source, and losing the context behind a result. GovDataHub puts a small, stable contract in front of those differences.
The Product Slice
- A catalog of ACS, business, and other configured U.S. Census datasets.
- Variable discovery and a query builder for geography, time, and predicates.
- Normalized JSON responses with CSV export and a browser-based explorer.
- Saved queries and repeatable catalog/refresh jobs for a working data workspace.
Engineering Signals
- FastAPI, async SQLAlchemy, Alembic, HTTPX, and Pydantic settings.
- Cache-aware results that retain source URL, fetch time, expiry, and checksum.
- Last-good metadata preservation and partial-failure reporting during refreshes.
- Accessible, responsive UI states for loading, errors, empty results, and overflow.
Commercial Reuse
The service is designed around U.S. Census statistical data, which is generally available for commercial reuse. The operational boundary is explicit: keep Census attribution and source links, follow API terms and rate limits, avoid implying agency endorsement, and review any third-party notices attached to a selected dataset.
Current Boundary
This is a private, working prototype rather than a public hosted endpoint. The next product decision is whether a narrowly scoped hosted API or an authenticated team workspace creates enough value to justify operating cost and upstream quota.