What we have validated, what we have not, and what a number means.
Linked state disclaimer permits copying/distribution for personal or informational use, with third-party rights caveat and no endorsement. Retained bounded public county facts; no map tiles, graphics or page layout served. Manual reviewed import only.
Reviewed public agency-linked county aggregate dashboard and as-is/no-endorsement policy. Private bounded capture, proposed attributed facts-only free informational release; no copied layout/media or implied license. Manual reviewed import only.
Reviewed public agency-linked county aggregate dashboard and as-is/no-endorsement policy. Private bounded capture, proposed attributed facts-only free informational release; no copied layout/media or implied license. Manual reviewed import only.
Agency-linked public aggregate residents by county and rash-onset year. Noncommercial public informational use/fair-use policy reviewed; facts-only proposed release, no dashboard/media reproduction or endorsement. Manual reviewed capture only.
Reviewed public agency-linked aggregate county-of-residence table. Facts-only local capture; no copied dashboard layout/media served. Kentucky privacy/legal notice reviewed; proposed free informational publication requires exact release review. No automated dashboard export or refresh.
CDC public-domain data; attribution and non-endorsement required. Robots checked again before each manual fetch.
Texas policy permits noncommercial copying without alteration, subject to exceptions. This local preview shows attributed aggregate facts; public reuse needs an intended-use review. No images or full pages are served. Robots gate required.
Documented metadata API; no articles or images copied. US relevance checked per item.
Public aggregate annual diagnosed residents only. Cross-year outbreak/exposure figures excluded. Utah informational-use terms; 10-second crawl delay.
Local attributed public aggregate facts from the agency-linked public dashboard and dated release. Browser capture retained privately; county assignment basis unspecified. Public intended-use review remains separate. Reviewed snapshot import only; no automatic dashboard scraping.
Public aggregate frequency query, confirmed cases by county of residence and report date. Local facts with attribution; retained evidence private. Reviewed bounded form snapshot only; public intended-use review remains separate.
The official report date describes the data. Retrieval records when the source bytes arrived; a successful validation or unchanged recheck is not a new report. Accepted snapshots and revisions are immutable. Unexplained decreases are retained for review, never converted into negative daily infections.
CDC totals keep U.S. residents and international visitors distinct. CDC national and jurisdiction values must come from one snapshot. New York City is a separate jurisdiction; New York State excludes it.
Latest state reports use their own publication dates and scopes. County assignment follows each source: residence when explicit, otherwise agency-published county with that limitation. These are not exposure sites. We never add mixed-date state values into the CDC total or infer missing counts from map colors.
Latest county reports are an independent overlay, including over the CDC state map. Available county reports keep their own dates and definitions, separate from CDC fills and totals. Unsupported years have no county circles. Circle area is proportional to positive reported cases: radius = k times the square root of count in CSS pixels. National k is 1.2 (0.9 below 600px map width); focused k is 2 for every state. The legend uses the same scale. The Census 2024 county Gazetteer supplies verified FIPS identifiers and interior points, separately from case provenance. These points are not patient or exposure addresses. Unlisted counties are absent, not zero; suppressed and unmatched records stay explicit in the county list.
Annual outbreak totals are cumulative, not a count of active outbreaks. Unexplained decreases are quarantined, with the prior validated snapshot retained. Failed refreshes retain the prior accepted snapshot with a visible warning; sources without accepted data remain unavailable. No demonstration cases are substituted.
Versioned Census boundaries are cached locally. CDC exact jurisdiction values reconcile to its resident total. Utah reports annual diagnosed residents; Maryland reports confirmed annual cases. These definitions and dates stay separate. Pennsylvania’s own adapter, population denominators and historical trends remain unavailable.