Provenance method
A receipt is only useful when it is generated from the data model.
DataByCounty uses a shared metric registry to connect rendered county facts to source agencies, source fields, vintages, transforms, confidence, and canonical destination pages across the ByCounty Network.
Stable metric IDs
Every receipt row is backed by a typed metric ID in the shared registry. Copy should not invent fields the pipeline does not produce.
Visible gaps
Null values stay visible. Missing source coverage, suppression, and unfinished transforms should be easier to find, not hidden behind averages.
Change-ready
The registry is structured for future patch notes: source update, methodology update, normalization update, or missing-data repair.
Confidence
Confidence is a source-quality label, not a judgment about the county.
High confidence usually means a direct county estimate from a documented public source. Medium confidence often means county aggregation, fallback matching, or derived scoring. Limited confidence flags values that are useful but need more context before they drive a decision.
| Label | Meaning |
|---|---|
| High | Direct source field or strongly documented county-level estimate. |
| Medium | Derived, aggregated, or transformed from public records with clear caveats. |
| Limited | Useful as context, but sensitive to coverage, model limits, or source gaps. |
What comes next
The registry can become a build-time manifest from the data pipeline. Once each pipeline run emits metric vintages and transform versions, DataByCounty can publish county patch notes showing what changed and why.
Inspect current sources