NYC Pool Inspections by Month: When Violations Peak

In NYC's public pool inspection data, routine inspections recorded at least one violation most often in January (60.6% of 325) and least often in August (42.5% of 1,306) and December (42.6% of 517). August also had more routine inspections than any other month, so it produced the most inspections with a violation in absolute terms. This is a read of inspection records from 2019 to 2025, not a measure of how safe any pool was.

Our first NYC post gave the overall shares. This one goes through the month, permit-type, borough and re-inspection tables and ends with a readiness calendar you can copy.

What is in the data?

The source is the DOHMH Pool Inspections dataset on NYC Open Data (3kfa-rvez), pulled 2026-10-21: 7,080 rows dated 2019-01-02 to 2025-12-31, covering 1,221 permits. We count "routine" as inspection type exactly "Routine Inspection" (5,585 rows). Violation shares are the share of routine inspections with at least one violation. Critical and public health hazard (PHH) shares use only rows where that count was recorded, 4,257 of the 5,585 routine rows. We publish totals and shares only, with no facility names.

What does the month-by-month table show?

Month	Routine rows	1+ violation	1+ critical	1+ PHH	Non-routine share of all rows
Jan	325	60.6%	26.0%	25.6%	26.0%
Feb	302	57.9%	18.7%	24.7%	32.6%
Mar	387	56.3%	22.4%	22.4%	24.3%
Apr	286	52.1%	18.8%	19.3%	30.8%
May	239	58.2%	26.5%	20.5%	29.1%
Jun	222	58.1%	18.2%	24.2%	32.5%
Jul	753	51.0%	20.4%	11.8%	17.2%
Aug	1,306	42.5%	15.9%	12.8%	14.9%
Sep	388	51.8%	19.2%	18.3%	23.0%
Oct	420	52.1%	20.2%	14.8%	18.8%
Nov	440	48.9%	20.1%	14.7%	15.9%
Dec	517	42.6%	22.4%	13.3%	16.1%

The last column is our arithmetic: all rows in the month minus routine rows, divided by all rows (for example August, 1,534 - 1,306 = 228, and 228 / 1,534 = 14.9%).

Four things stand out:

• Two halves. January to June ran from 52.1% to 60.6% with any violation. July to December ran from 42.5% to 52.1%. PHH shares split the same way: 19.3% to 25.6% in January to June, 11.8% to 18.3% in July to December.
• Volume is lopsided. July and August hold 2,059 of 5,585 routine rows (36.9%). April to June hold 747 (13.4%). August has 5.5 times May's routine rows (1,306 / 239).
• Count versus share. August: 1,306 x 42.5% is about 555 inspections with a violation. January: 325 x 60.6% is about 197. A lower share in a busy month still means more inspections to answer for. These products use rounded shares, so they are approximate.
• Follow-up tends to move with the share. Months with higher violation shares mostly had higher non-routine shares: February 57.9% and 32.6%, June 58.1% and 32.5%, against August 42.5% and 14.9%. This is a pattern, not proof. Non-routine rows include complaints, a re-inspection can land in a later month than the finding, and spring months have few routine rows, so a steady number of follow-ups makes a larger share.

Why? The summary can't say. It does not cross month with permit type, so we cannot show that winter rows are mostly indoor pools or that August is mostly outdoor. Indoor permits hold 4,800 of 7,080 rows overall and 3,625 of 5,585 routine rows (64.9%), so a mix effect is plausible, but it is a hypothesis. Other candidates: pools reopening after a closed period, how inspectors schedule visits, staff turnover. Check them against your own records.

How much can you trust one month's share?

Not as much as the table's precision suggests. NIST's e-Handbook gives the Wilson interval for a proportion: (p + z^2/2n +/- z x sqrt(p(1-p)/n + z^2/4n^2)) / (1 + z^2/n), with z = 1.96 for 95%. Applied to rows from the table:

Group	Share	Routine rows	95% Wilson interval
Jan	60.6%	325	55.2% to 65.8%
May	58.2%	239	51.9% to 64.3%
Aug	42.5%	1,306	39.8% to 45.2%
Dec	42.6%	517	38.4% to 46.9%

This assumes every inspection is independent. They are not: the same permits are inspected again and again, so the true uncertainty is wider. Even so, January and December do not overlap, so "winter" is not one block.

Year matters too. Each month pools up to seven years, and the year-to-year spread is larger than the month spread:

Year	Routine rows	1+ violation	1+ critical	1+ PHH	Rows with critical count recorded
2019	911	62.3%	28.0%	16.6%	911
2020	389	37.8%	10.3%	13.9%	389
2021	704	40.8%	12.5%	8.7%	704
2022	829	51.9%	35.8%	25.6%	430
2023	887	56.5%	23.4%	20.6%	501
2024	905	49.8%	9.3%	13.5%	905
2025	960	43.4%	27.1%	25.9%	417

Two caveats. First, critical and PHH subtotals are blank on 571 rows (all inspection types) in 2022, 562 in 2023 and 788 in 2025 (789 for critical in 2025), and on none in 2019 to 2021 or 2024, so those years' shares rest on fewer rows. A blank is a missing count, not a clean result. Second, 2020 has 483 inspections against 1,267 in 2019 and none from April to June, so April to June cover six years while other months cover seven. The month-by-year split is not in the summary, so we cannot separate the two effects.

What changes between indoor and outdoor permits?

Permit type	Routine rows	1+ violation	1+ critical	1+ PHH	Average violations	Median
Indoor	3,625	53.1%	21.4%	19.0%	1.03	1
Outdoor	1,960	44.7%	17.3%	12.0%	0.83	0

The gaps are 8.4 points for any violation, 4.1 for critical and 7.0 for PHH (19.0 / 12.0 is about 1.6 times). Hypotheses include more equipment and air handling indoors, year-round operation, different inspection timing, and different permit mix. The dataset holds none of those, so we assert no cause.

Do boroughs differ?

Borough	Routine rows	1+ violation	1+ PHH
Brooklyn	1,096	58.9%	13.4%
Queens	824	55.8%	18.5%
Bronx	448	50.7%	16.3%
Manhattan	2,710	48.4%	19.1%
Staten Island	507	31.2%	6.7%

Staten Island has 507 routine rows against Manhattan's 2,710, so its share moves more with a few permits. By the Wilson formula above, its 31.2% spans 27.3% to 35.4% even if inspections were independent, which they are not. Do not read a borough ranking as a ranking of operators.

How heavy is the re-inspection burden?

Of 1,221 permits, 482 (39.5%) had at least one re-inspection and 610 (50.0%) had at least one non-routine inspection. Re-inspection means any type containing "re-inspection", 992 rows against 5,585 routine rows, about 18 per 100 routine inspections (our arithmetic).

A repeat PHH is a permit with PHH of at least 1 on two or more routine inspections. Counting blanks as none, that is 163 of 1,003 permits with two or more routine inspections (16.3%, a lower bound). Using only the 961 permits with PHH recorded on two or more, it is 17.0%. Same permit and same day complaint plus routine pairs, 118 of 137 same-day groups, count as separate rows. The dataset has no violation descriptions, so a repeat does not mean the same problem.

How do you turn this into a readiness calendar?

Swap in your own season dates and local busy months.

Window	Task	Evidence to pull	Owner
4 to 6 weeks before an outdoor pool opens	Full record review	The 17 daily record items in MAHC 6.4.1.6, equipment dates, certificates	Area supervisor
Two weeks before local busy months	Spot-check the busiest sites	Last 30 days of logs, open repairs, staff certificate dates	Regional manager
Every quarter for year-round indoor pools	Same review, plus air and equipment notes	Logs, repair history, prior inspection reports	Area supervisor
Within a day of any inspection	Re-inspection follow-up log	Finding, fix, who, date, retest, photo	Site head guard
After each inspection	Close the loop	Compare your report to your last review	Ops manager

Make a re-inspection log row for every finding: date, type, finding, fix, owner, closed date. Keep the original finding visible after it is closed.

Example (made-up numbers): a company runs 30 pools, 12 indoor all year and 18 outdoor. A record review takes 1.5 hours per pool. Pre-season for the outdoor pools is 18 x 1.5 = 27 hours, or 9 hours each across 3 supervisors. The quarterly indoor review is 12 x 1.5 x 4 = 72 hours a year. Put the 27 hours before the opening dates, not in the same weeks as their busiest inspections.

How often inspectors come is up to your jurisdiction; MAHC 6.6.1.3 (check what your state adopted) lets inspection frequency be adjusted based on risk. Poolean builds daily and monthly logs from each pool's readings, visits and closures, and flags pools that need attention before an inspector arrives, with the reasons. Book a 30-minute call to see it. Also see inspection preparation and inspection risk across sites.

What can't this data tell you?

• It is counts only. It has no violation descriptions, so we cannot say which problems were common.
• It is inspection records, not a measure of pool safety.
• It is NYC only. Do not apply NYC shares to another county.
• Month, year and permit type are mixed together in the summary tables.

The short version

• January had the highest share of routine inspections with a violation (60.6%); August and December the lowest (42.5% and 42.6%).
• August carries 1,306 routine rows, 5.5 times May, so volume peaks when the share is low.
• Indoor permits: 53.1% any violation, 19.0% PHH. Outdoor: 44.7% and 12.0%.
• 39.5% of permits had a re-inspection. Track findings to a checked fix.
• Build the same table from your own reports with denominators shown.