How We Analyze Water

We do not run our own lab tests. Where a source does not cover a place or a substance, the report says “Not measured”. All sources and their limits →

1. STEP 1

   Collect

   Download the public release (UCMR 5, SDWIS, service-line inventory, utility report).

2. STEP 2

   Keep the context

   Publisher, sampling date, location, unit, qualifier and limit stay attached to each result.

3. STEP 3

   Validate

   Unreadable or contradictory results are kept as visible gaps, not dropped or guessed.

4. STEP 4

   Compare

   Each usable result is compared only with its cited government yardstick.

5. STEP 5

   Match

   A finding that calls for treatment is matched to a technology and the certification to check.

The Purity Score (0–100) summarises how a water system's measured results and compliance history compare with government limits. The score belongs to the water system. A city page shows the score of the system that serves most homes there, and lists other systems with their own scores. We never average scores across systems.

Each result is read against the first yardstick that exists, in this order: the federal MCL or action level, then an EPA health advisory or health reference level, then the strictest US state MCL. Only government sources count, and every yardstick is cited. A contaminant without one is shown but not graded; we never invent a limit. Lithium is shown with its EPA health reference level (10 µg/L, a screening value) but is not graded. Three steps, no hidden inputs:

1. 1

   Score each graded contaminant

   • Not detected 100
   • Below the limit 100 − 40 × (value ÷ limit) from 100 down to 60
   • Above the limit 50 − 50 × (value − limit) ÷ limit 50 at the limit, 0 at twice it
2. 2

   Each contaminant subject takes its lowest-scoring result

   PFAS, metals (lead and copper), disinfection by-products and other. One high reading is not averaged away by clean ones. The track record is not a lowest-scoring result: it starts at 100 and loses points for each health-based violation (below).

3. 3

   Weighted average of measured subjects

   PFAS 30% · metals 30% · disinfection by-products 20% · other 10% · track record 10%. Subjects without a usable result are left out and the weights re-scale, so a missing test never counts as clean. Each contaminant score and the overall are rounded half up (74.5 becomes 75). Hardness is never scored.

When a subject has a result above its limit, the band says so, for example “Fair (PFAS above federal limits)”, or “above health guidelines” when the yardstick is not a federal limit. The label never hides the finding behind it.

The track record comes from national EPA compliance data, so it is always measured. Missing data lowers confidence, never the score and never raises it. Each report lists the subjects still pending a source.

A PFAS compound with none of these yardsticks is shown with its measured value but not graded. PFAS mixture rules (the federal Hazard Index and state sum-of-PFAS limits) are not part of the score yet; each compound is read on its own.

A filter qualifies only with a certification we checked in the NSF, WQA or IAPMO listing for that exact contaminant. Merchant offers are attached afterwards and never change which filters qualify. Try the filter finder →

A page is offered to search engines only when it carries a sourced fact no other page has. Thin pages stay available to readers but out of the index.