Available now

US air quality data: daily AQI by county and metro area, and pollutant readings by monitor

EPA Air Quality System data from the AirData files: the daily AQI for every county and metro area with a monitor since 1980, daily pollutant readings at about 10,600 monitors since 2010 and annual summaries for about 1,200 parameters. 15 tables, about 55 million rows.

Coverage: Daily AQI for about 1,500 counties and 690 metro and micro areas, 1980 to 2026. Daily monitor summaries for ozone, PM2.5, PM10, NO2, SO2 and CO, 2010 to 2026. Annual monitor summaries, 1980 to 2026. About 21,000 monitoring sites. 2026 is partial and not yet certified: it runs to EPA’s latest AirData update, which EPA posts each June and December. Counties without a monitor have no rows.

Why pay for public data

EPA publishes AirData as 356 zip files in six layouts, and its annual county AQI file carries county names but no FIPS codes. This package types and keys every file on zero-padded state and county FIPS codes and SITE_ID, so air quality joins to Census and CDC county data, and adds reference tables and a flag for the PM2.5 rows EPA says to exclude.

How it is built

  1. 1

    Pull the EPA AirData daily and annual monitor summaries, the daily and annual AQI files by county and metro area, and the AQS site, monitor and code tables.

  2. 2

    Type every column, key the tables on STATE_FIPS, COUNTY_FIPS and SITE_ID, match the annual county AQI file to FIPS codes by county name, and check row counts against EPA’s own file list.

  3. 3

    Flag the Teledyne T640 and T640X PM2.5 rows that EPA’s 2024 data advisory says to exclude, and remove the name and address of about 220 old sites at private residences. Refreshed after EPA’s June and December AirData updates.

What is in each record

FieldNotes
STATE_FIPS, COUNTY_FIPS, SITE_IDZero-padded FIPS codes. SITE_ID joins the monitor tables to SITE
AQI_DATE, AQI, AQI_CATEGORY, DEFINING_PARAMETER, DEFINING_SITE_ID, NUMBER_OF_SITES_REPORTINGDAILY_AQI_COUNTY and DAILY_AQI_CBSA: about 12 million and 6.4 million rows
GOOD_DAYS, UNHEALTHY_DAYS, MAX_AQI, MEDIAN_AQI, DAYS_OZONE, DAYS_PM25ANNUAL_AQI_COUNTY and ANNUAL_AQI_CBSA: days in each AQI category and days each pollutant set the AQI
PARAMETER_CODE, DATE_LOCAL, ARITHMETIC_MEAN, FIRST_MAX_VALUE, FIRST_MAX_HOUR, AQIDAILY_MONITOR_SUMMARY: about 33 million rows
SAMPLE_DURATION, POLLUTANT_STANDARD, EVENT_TYPEEPA builds several rows per monitor and day. Filter on these three
PM25_ADVISORY_EXCLUDERows that EPA’s 2024 PM2.5 data advisory says to exclude
YEAR, PERCENTILE_98, PRIMARY_EXCEEDANCE_COUNT, COMPLETENESS_INDICATOR, CERTIFICATION_INDICATORANNUAL_MONITOR_SUMMARY: about 2.9 million rows, including air toxics, VOCs, PM speciation and meteorology
LATITUDE, LONGITUDE, LAND_USE, LOCATION_SETTING, SITE_ESTABLISHED_DATE, SITE_CLOSED_DATESITE: about 21,000 sites, with county FIPS and CBSA
MONITOR, PARAMETER, METHOD, POLLUTANT_STANDARD, AQI_BREAKPOINTAbout 370,000 monitors and the AQS code tables
COUNTY, CBSA, DATA_DICTIONARYLookups, and one row for each of the 272 columns

Who uses it

  • Public-health and epidemiology researchers building exposure features
  • Insurers, real estate and site-selection analysts
  • Demand planners for air purifiers, HVAC and allergy products

Pricing

$399 a year, including each twice-yearly refresh.

No personal contact data is collected or sold in any QuestionPunk dataset.

Questions

Is this data not free from the government?
Yes. The source files are public and we link to them. You pay for the cleaning, the joins, the documented schema and the refresh, not for the data itself.
How is it delivered?
As Parquet or CSV files, by download link or to your own cloud bucket. The same tables load into Snowflake with the scripts we include.
Can I see the data before paying?
Yes. Ask for a sample and we send the Vermont rows from every table, plus the full data dictionary.
How current is it?
2026 rows run to EPA’s latest AirData update, which EPA posts each June and December, and are not yet certified. 2025 is a full year. Agencies certify each year by May 1 of the next year, and CERTIFICATION_INDICATOR shows the status of every annual row.
Does it include hourly data or official design values?
No. Hourly and 8-hour files, daily monitor summaries before 2010, daily files for toxics, speciation, lead and weather, AirNow data and official design values are not included. The annual summaries for those parameters are.
Does it join to Census and CDC county data?
Yes, on COUNTY_FIPS, except in Connecticut: EPA uses the legacy counties, while recent Census and CDC releases use planning regions. A few border monitors carry AQS codes for Mexico and Canada.
Does it include personal data?
No. Every row is a measurement, a site or a code. For about 220 old sites at private residences, the site name and address are removed and the coordinates are rounded.
What is the license?
EPA’s AirData FAQ says the ambient monitoring data in AQS are public domain. EPA does not endorse this product.

Tell us what you need

Geography, size and fields. A two-line email is enough, and a person replies.

Request a sample

More from the catalog