Permit Pulse: New York City — week of 2026-08-17
Building permits issued 2026-08-17 – 2026-08-23. 844 citywide (-56.9% against a 1957/week trailing baseline, z -3.1) across 219 ZIP codes.
The city as a whole moved abnormally this week. New York approved 844 permits against a trailing average of 1,957 — a 56.9% fall, 3.1 sigma below its own year. Every ranking below should be read in that light: on a week like this the dominant fact is citywide, not local, and a ZIP that merely held its usual count will look like it gained ground simply because everything around it stopped.
Heating up
| # | ZIP code | Permits | Trailing avg | Change | Share z | Own z |
|---|---|---|---|---|---|---|
| 1 | 11367 | 20 | 7.35 | +172.1% | 5.1 | 1.6 |
11367 (Flushing/Kew Gardens Hills, Queens) issued 20 permits against a trailing average of 7.35 — nearly triple its normal week. It is the only ZIP in the city that gained on both measures this week, and that is the point: its share z of 5.1 is enormous partly because the rest of New York went quiet, but its own count is genuinely elevated too, at 1.6 sigma above its own history.
Share of city: 2.37% against a trailing 0.37% (sd 0.39, sigma used 0.392) = share z 5.1 Own count: 20 against 7.35 (sigma 7.911) = z 1.6
Both bars had to be cleared. Ten other ZIP codes scored between 3.4 and 5.1 on share this week without their own counts moving at all — they had simply held steady while the city fell by more than half. None of them appears above, because a neighborhood that did not change is not a neighborhood that surged.
Cooling off
| # | ZIP code | Permits | Trailing avg | Change | Share z | Own z |
|---|---|---|---|---|---|---|
| 1 | 10128 | 1 | 19.54 | -94.9% | -2.3 | -2.3 |
10128 (Upper East Side) issued a single permit against a trailing average of 19.54. On a week the whole city halved, this is one of the few areas that fell far enough on both measures to be distinguishable from the citywide slowdown.
What was permitted
- General Construction — 143 (16.9%)
- Sidewalk Shed — 119 (14.1%)
- Solar — 92 (10.9%)
- Plumbing — 89 (10.5%)
- Supported Scaffold — 73 (8.6%)
- Mechanical Systems — 72 (8.5%)
What this week does not cover
No ZIP code was excluded for thin history. 64 of the 219 ZIPs with data were held out of the ranking for averaging under 3 permits a week. 153 were scored and moved less than the bar in one or both signals. Neither table hit its limit, so nothing was cut. The trailing-year query did not hit the portal's row ceiling.
New York publishes estimated_job_costs as a text field on this dataset, which cannot be aggregated, so this edition reports no construction value. That is an absence of usable data, not a week without value.
How this is built. Source: New York City open data portal — building permits (https://data.cityofnewyork.us/d/rbx6-tga4). Every permit approved in the week is counted by ZIP code and compared with that ZIP's own trailing 52 weeks.
Two independent signals must agree before an area is called anything. The share signal asks whether the area took an unusual slice of the city's week; the own signal asks whether the area had an unusual week for itself. A surge needs share z ≥ 2 AND own z ≥ 1; a cooldown needs the mirror. This week is exactly why: with the city down 56.9%, share alone would have reported eleven surging neighborhoods, ten of which had not moved.
Excluded from ranking: areas with fewer than 26 weeks of history (reported as new), and areas averaging under 3 permits a week. Standard deviations are floored — a count at Poisson, a share at what one permit is worth — so a flat history cannot mint a large z. At most 10 areas appear in each table and the remainder is counted.
Permit data is public record from the city's own portal, published through 2026-08-28. Rubric permit-pulse/v1. Not real-estate advice.