Data Trust Center (formerly Data Review Hub)
What's fresh, where it comes from, and what's estimated. Unified inventory of all COHO Analytics data sources — freshness status, source-level QA checks, and automated discovery of new data files.
Data Health Overview
About This Page
This is the single place to answer "can I trust this data?" — Sources shows where every dataset comes from, Quality shows current freshness status, and QA Coverage shows whether the data passed automated structural checks and whether your place uses precise or estimated methodology. Browser-side scans detect new or unregistered public data files and surface them in the Pending Discovery tab for administrator review.
Raw Files
More detailed, maintainer-oriented views (API key setup, raw pipeline logs) are linked at the bottom of this page.
Data Sources Inventory
Loading sources…
Data Quality & Freshness
Checking…Detail Table
| Dataset | Status | Records | Cache age | Notes |
|---|---|---|---|---|
| Loading… | ||||
QA Coverage
Beyond per-source freshness (the Quality tab), this checks whether the site's data passes structural and plausibility checks, and whether each place's numbers come from a precise tract-level calculation or a broader county-level estimate.
What data is missing or estimated?
Every dataset runs through five automated checks — shape, row counts, numeric ranges, freshness against its own update schedule, and cross-source agreement — before it's considered trustworthy. See CONTRIBUTING.md for the full technical definition of each layer.
Status: loading…
Is my place using precise or estimated data?
Housing-need figures use two methodologies depending on the place: tract-aggregated (a TIGER 2024 spatial join — accurate for cross-county jurisdictions like Erie, Aurora, or Longmont) and primary-county fallback (used when a place isn't in TIGER coverage, so it inherits its containing county's rates). This panel shows which bucket every place falls into so the provenance is auditable rather than implied.
Pending Discovery
Public files detected but not yet registered in the browser-side inventory. Approve to download a manifest patch file.Run a scan to detect unregistered data sources.
Admin & Configuration
Manual Discovery Scan
Probe all known data file paths against the registered manifest to surface new or unregistered sources.
Discovery Configuration
Discovery configuration is maintainer-only in the public build; public visitors can run the scan against the shipped file manifest without loading private audit artifacts.
Public manifest onlyGitHub Actions Workflow
The data-source-monitoring workflow runs daily to:
- Scan for new/unregistered data files
- Check data freshness against expected update windows
- Generate discovery reports (JSON + Markdown)
- Create GitHub issues for new unprocessed sources
Workflow file: .github/workflows/data-source-monitoring.yml
Probed File Paths
The following paths are probed during each discovery scan:
- Loading…