Methodology

MyTapWater.us holds the data on this site to the highest possible standard. If we cannot verify a value, we leave it out. If a number is uncertain, we say so. Every page footer lists when the underlying dataset was last refreshed.

Data sources

1. EPA SDWIS via ECHO bulk downloads. Quarterly downloads from echo.epa.gov. Includes the universe of public water systems, violations, enforcement actions, and sampling.
2. EPA Envirofacts REST API. Used for spot-checks and cross-verification against the bulk download. Envirofacts documentation.
3. EPA UCMR (Unregulated Contaminant Monitoring Rule). Annual data on emerging contaminants including PFAS, lithium, and chromium-6.
4. State Lead Service Line Inventories. Required by the EPA's Lead and Copper Rule Revisions; deadlines and formats vary by state. We ingest each state separately.
5. U.S. Census TIGER/Line shapefiles and HUD ZIP crosswalk. For mapping ZIP codes to counties and places.

How we verify (two-stage gate)

Every quarterly release passes two automated checks before any page is updated.

Stage A: internal consistency

• Row-count delta against the prior canonical dataset. Any movement greater than 10% in a table or 5% in a state's PWS count fails the check.
• Required-field null-rate check. Any increase greater than two percentage points fails.
• Referential integrity. Every PWS ID referenced in a violation or sample row must exist in the PWS table.
• Numeric sanity. Concentrations must be non-negative and within physically plausible bounds.
• Unit consistency. Mixed mg/L and μg/L within a contaminant fails.

Stage B: external cross-verification

For a deterministic random sample of 25 public water systems per state per release, the script re-fetches the same record via the EPA Envirofacts REST API and diffs the canonical fields against the bulk download. Any discrepancy above tolerance fails the build.

What you will not see on this site

Health-risk classification

Where MyTapWater.us displays a safe / caution / concern badge for a contaminant, the threshold comes from the EPA's published Maximum Contaminant Level (MCL), Maximum Contaminant Level Goal (MCLG), or Health Advisory Level (HAL). We do not invent thresholds. If the EPA has not published one for a given contaminant, we display the measured value without a badge.

Known limitations

• ZIP to public water system mapping is approximate. The EPA does not publish authoritative service area boundaries for most systems. Where a state has published shapefiles, we use them. Otherwise we use the system's administrative address as a centroid plus population-served as a heuristic. Ambiguous ZIPs show every candidate system with a confidence label rather than auto-picking one.
• State LSL inventories vary in completeness and recency. Some states have not yet posted theirs. Others use different category labels. Where data is missing, the page says so.
• SDWIS quarterly downloads lag real-time enforcement. We re-pull every quarter and cross-verify against the live EPA API for the sample, but a violation issued today will not appear here until the next refresh.

Update cadence

• EPA SDWIS: pulled quarterly (January, April, July, October)
• EPA UCMR: pulled annually
• State LSL: pulled quarterly per state, with the index rebuilt whenever a state updates
• ZIP and Census boundaries: pulled annually unless a boundary changes sooner

Corrections policy

If you find a value on MyTapWater.us that disagrees with a primary source (your utility's Consumer Confidence Report, the state inventory, the EPA ECHO viewer, etc.), please email hello@mytapwater.us with the page URL, the specific value, and the source you consulted. We investigate within five business days and either correct the value, annotate the discrepancy with a note, or explain why our value is the one we will keep.

Affiliated with the EPA?

No. MyTapWater.us is independent of the EPA, of any state agency, and of any water utility. We use public data published by these agencies and cite each dataset on the page where it appears.