Public data, made legible.
The best neighborhood data in the country is already public. It is just impossible to use.
Six agencies publish the metrics that decide where people live, each in its own format, cadence, and geography. Atlasblock does the joining, the normalizing, and the caching, and hands back one honest number per category.
Public data, made usable.
The federal government publishes remarkable open data. It is scattered across a dozen formats and portals. We do the joining so you do not have to.
One number, honest provenance.
Every rating traces back to a named source and a documented formula. No black box, no proprietary blend you cannot audit.
Fast enough to score a feed.
A cached lookup returns in tens of milliseconds, so you can rate every listing in a search result without blocking the page.
One zip in. Eight ratings, raw metrics, and a written profile out.
A weekend crosswalk
It started as a script to join EPA walkability scores to zip codes for one house hunt.
More ratings
Crime, schools, and density joined walkability, each normalized to a single honest scale; climate became an objective facts block rather than a score.
The cache
Aggressive per-zip caching brought first-byte latency under 40 milliseconds and made feed scoring practical.
Public API
Live and test keys, credit-based pricing, and a documented response schema opened the data to any builder.
Dataset 2026-Q2
The current release refreshes federal sources and expands coverage to 33,142 zip codes.
A small team of engineers and civic-data nerds who got tired of copying numbers out of government spreadsheets. We keep the surface small, the provenance visible, and the cache warm.
