@Ondřej-Šutera Thanks. I suspect it's still very hard to get right, the localized summarization and presentation of this complex data. I would hope for Windy to report at least 50th percentile, or preferable 90th percentile AQI averaged over a small region. I'm bothered that Windy happily reports inaccurate data, that seems reasonable on most days, but on days with poor air quality when the numbers matter most, Windy's PM2.5 numbers turn out to be terribly misleading.
I'm concerned that since Wednesday, PurpleAir has reported "PM2.5 AQI" of 200 for the entire inner Bay Area from San Jose to Berkeley to San Francisco, never (when I've checked) breaking below 120 until this afternoon, but in that same period, I have not seen a Windy-reported PM2.5 point sample above 10ug/m^3. Based on Wikipedia's AQI article these numbers are entirely incompatible, both for US AQI scales and Europe's CAQI/EAQI scales. I don't see Windy's PM2.5 reporting for locations in Berkeley, San Francisco, or San Jose correlate with data available on Purple Air. I don't see Windy reporting unhealthy AQI values here at all, when is unhelpful.
Some less testimonial evidence:
PurpleAir history for a Berkeley sensor:
us-epa-pm25-aqi.png
Current mismatch, showing the consistent, erroneous PM2.5 concentration (50th/90th/95th percentile would be reasonable) reported by Windy for all of the Bay Area.
Screenshot_20230922_173209.png
My experience using this, is like hearing: 3.6 roentgen per hour, not great, not terrible; when that is not the state of the world.
