NYC Pool Inspection Data: How Often Violations Were Found

Half of NYC's routine pool inspections found at least one violation: 50.2% of 5,585. Indoor pools did worse than outdoor ones, and August was the busiest month for inspectors. The counts show patterns. They don't explain them, and they say nothing about whether a given pool is safe today.

Our source is the NYC Pool Inspections dataset on NYC Open Data, published by the NYC Department of Health and Mental Hygiene and pulled October 21, 2026. It contains 7,080 inspection records dated 2019-01-02 to 2025-12-31, covering 1,221 permits. NYC requires a permit to build or operate any indoor or outdoor pool, and one permit can have many inspections, so keep the two counts apart.

How often did routine inspections record violations?

We kept routine inspections separate from complaint and return visits so those wouldn't skew the result. Across routine inspections, the average violation count was 0.96 and the median was 1.

The dataset also tags violations as “critical” or “public health hazard.” We use those labels as recorded; there's no description of what actually happened.

Among routine inspections with the relevant count recorded, 20.0% had at least one critical violation and 16.7% had at least one public health hazard. Each percentage uses 4,257 routine records. Some category counts are missing, which is why the base is smaller than 5,585.

How did indoor and outdoor inspections compare?

Indoor and outdoor come from the permit type.

Permit type	Routine inspection records	Share with at least one violation
Indoor	3,625	53.1%
Outdoor	1,960	44.7%

That doesn't prove being indoors causes more violations. The two groups may differ in ways this summary can't see, and the same permit shows up in many inspections. Use the gap as a prompt to look at your own indoor sites, not as a rule that every indoor pool is riskier.

When were inspections busiest?

August had 1,534 inspection records, followed by July with 909 and December with 616. Those are all inspection types across all years, not violation rates. In fact, August had the lowest routine violation share of any month: 42.5% of 1,306 routine inspections, against 60.6% of 325 in January.

The monthly totals also combine years with different amounts of inspection activity. There were 1,267 inspection records in 2019 and 483 in 2020. There are no inspections at all from April through June 2020; the dataset page says pools were closed by order or chose not to open during the COVID-19 emergency, and inspections were paused or changed.

August's peak won't tell you when your next inspection is. It's a reason to keep records ready all season instead of scrambling when the inspector shows up.

Did hazard findings happen again under the same permit?

We counted a repeat when the same permit had a public health hazard on at least two routine inspections. It may not be the same problem twice; the dataset has no descriptions to check.

Among 961 permits with hazard counts recorded on at least two routine inspections, 163 met this repeat definition. The share was 17.0%.

Using all 1,003 permits with at least two routine inspections gives 16.3%. That treats blank hazard counts as no hazard, so it's a lower bound.

Follow-up visits were common too. 482 of the 1,221 permits (39.5%) had at least one re-inspection, and 610 (50.0%) had at least one visit that was not routine.

How we did this

We used the public dataset, not individual facility files, and counted rows in Python grouped by inspection type, date and permit type.

“Routine” means the inspection type was exactly “Routine Inspection.” Violation shares count routine inspections with at least one recorded violation. Critical and hazard shares use only records where that category's count was present. Percentages are rounded to one decimal place.

Routine and complaint records for the same permit on the same day stay separate. Repeat findings are grouped by permit. We publish only totals and shares, with no facility names or addresses.

What this data cannot tell you

• It has counts only, with no violation descriptions. We cannot say which problems were most common or whether a later finding was the same issue.
• Hazard counts are blank in 2022, 2023 and 2025. Across all inspection types, 1,921 hazard counts are missing. Critical and general category counts also have gaps. Blank does not mean clear.
• It covers NYC only. This is not a national rate, and other health departments inspect against their own codes.
• Correlation is not cause. Results can differ by month or permit type without either causing it. These records don't prove that software, staffing or a particular practice changed outcomes.
• It is not a current safety check. Old counts don't replace a look at present conditions or your health department's directions.

How do you run the same review on your own pools?

Poolean flags sites that need attention before an inspection, with the reasons. To check each one yourself:

1. List every inspection for each pool, with date and type (routine, complaint, re-inspection).
2. Mark each one as "violation found", "no violation", or "result missing". Do not count a missing result as clean.
3. For each violation, write the response, who handled it, and when someone checked the fix.
4. Flag any pool with a public health hazard on two or more routine inspections, like the repeat count above.
5. Keep the original finding visible even after the follow-up is done.

Start with our guides to preparing inspection records, building a water test log and reviewing inspection records across sites. You can also book a call to discuss organizing your pool records.

The short version

• Count inspections separately from permits when reviewing your logs.
• Show missing entries clearly instead of treating them as good results.
• Use patterns to choose what to review, not to claim a cause.
• Connect findings to actions and checked follow-up in your own records.