Can you trust the US air‑sensor network's data?

Why the checks are built this way

OpenAQ can ingest readings out of chronological order, so a feed can look briefly stale or gappy when the lag is upstream, not at the sensor. Two choices absorb that: completeness is measured over afixed trailing 30-day window rather than raw last-seen, so it reads the sensor's reporting health, not OpenAQ's ingestion lag; and staleness keeps a deliberately conservative 24-hourbar so ordinary lag never trips it. Windowed completeness and plausibility both come from one bounded per-sensor daily-aggregate call — no dense measurement pulls — so the whole panel scores on free rails.

The honest denominator

Today's run scored 5465 sensors:5465 plotted on the map and 0scored but unmappable (no coordinates from the provider — counted in every aggregate, just not drawn). A further 71were held out of scope before scoring — 71 Kentucky/Louisville and 0 whose provider license disallows redistribution — so they never touch the numbers above.

Panel: US PM2.5 sensors. Kentucky/Louisville sensors are held out as a conflict-of-interest precaution — keeping the author at arm's length from any local network. Two richer checks (flatline detection and cross-sensor consistency) are deferred to a later version.