Air & Water Score by ZIP Code

How this is calculated

Two independent scores, each from a different EPA dataset, combined into one. The composite only appears when both halves exist — where a county has no air monitor you get the water score alone, labelled as such.

Air

From EPA AirData's annual AQI by county file for 2025, the latest complete year. The score is anchored on the 90th percentile AQI — the level the county stays below on 90% of measured days. EPA publishes that figure precisely because it resists the single bad day that a maximum is hostage to.

air = 100 − (90th percentile AQI ÷ 2), floored at 0

The divisor maps the official AQI breakpoints onto the score: an AQI of 50 (the top of “Good”) gives 75, 100 (top of “Moderate”) gives 50, 150 (top of “Unhealthy for sensitive groups”) gives 25, and 200 or worse gives 0. Where a county reports fewer than 300 measured days we label the year partial rather than treating it as complete.

Water

From the EPA SDWIS/ECHO bulk download (2026Q2), covering every active community water system on record as serving the county — 49,378 systems nationally. Violations are counted over the last 5 years, and EPA's own status field does the work: a violation still open is a different fact from one that was fixed.

penalty per system = 25×(open health-based)
                   + 8×(resolved health-based)
                   + 4×(open monitoring/reporting)
                   + 1×(resolved monitoring/reporting)
                   capped at 100

water = 100 − (average penalty, weighted by people served)

Weighting by population served stops a 25-connection system outvoting a city utility; the per-system cap stops one failing system zeroing an otherwise sound county. Those four coefficients are our judgement, not EPA's, which is why the tool exposes them as inputs — change them and the score recalculates from the same unchanged counts.

Composite

score = 50% × air + 50% × water

Matching a ZIP code to the data

Neither dataset is published by ZIP code. SDWIS has a ZIP field, but states almost never populate it — of 578,199 service-area records, only 7,155 carry a ZIP, covering 1.7% of the country. Both halves are therefore matched by county, using the Census 2020 ZCTA-to-county relationship file. Where a ZIP straddles two or more counties — about 30% of them do — we use the county holding the largest share of its land area, and name that county on the result so the choice is visible.

Coverage, stated plainly

• Water: 99% of ZIP codes (98% of the population), across 3,169 counties.
• Air: 52% of ZIP codes (77% of the population), across 985 counties with monitors. Most people are covered even though half of ZIP codes are not, because monitors sit where people are.
• PO-box and single-address ZIP codes have no Census tabulation area and return no result. The dataset covers 33,791 ZIP codes.

What this cannot tell you

• Violations are compliance records, not measurements of your tap.Lead service lines on the customer's side, private wells, and anything not sampled under the Safe Drinking Water Act are invisible here.
• A score describes a county and a utility, not a house. Air quality varies within a county — next to a freeway is not the same as five miles away.
• We never interpolate. A missing half is shown as missing.

Sources

• EPA AirData — annual AQI by county (2025)
• EPA ECHO — SDWA (SDWIS) bulk data download (2026Q2)
• US Census Bureau — 2020 ZCTA to county relationship file
• EPA AQS — state and county code table

Last reviewed: September 2026 · Data in use: SDWIS 2026Q2, AirData 2025 · Dataset built 2026-09-02