Skip to content
Docs

Subway wait assessment

The share of gaps between subway trains that meet the MTA's standard, by month, line and period, since January 2015. The MTA's own figure, copied from data.ny.gov.

Rows
12,812
Size
158 KB
Covers
January 2015 to last month
Updated
Nightly (the MTA publishes monthly)
Source
MTA, on data.ny.gov (subway wait assessment)
Fileshttps://data.transitlab.nyc/v1/subway_wait_assessment.parquetOne file. In the database as transitlab.subway_wait_assessment.

Rows and size from catalog.json, 2026-10-08.

One line in one month, in the peak or off-peak hours, on weekdays or weekends. The MTA checks the gap between trains at chosen stops (timepoints). A gap passes if it is no more than 25% longer than the scheduled gap. The MTA calls this a legacy measure.

We changed two things from the MTA’s file: month is a date and day_type says weekday or weekend instead of 1 or 2. The figures are the MTA’s.

SQL
-- Share of good gaps by line, weekday peak, 2025
SELECT line,
round(sum(num_timepoints_passing_wait_assessment) / sum(num_sched_timepoints), 3) AS share_good
FROM transitlab.subway_wait_assessment
WHERE day_type = 'weekday' AND period = 'peak' AND year(month) = 2025
GROUP BY line
ORDER BY share_good;
monthdate

First day of the month.

divisionvarchar

A DIVISION (numbered lines and the 42nd St shuttle) or B DIVISION (lettered lines).

linevarchar

Subway line. J and Z run as one line, JZ. Shuttles are S 42nd, S Fkln and S Rock. Some lines appear under more than one name, or in some years only: J and JZ, the shuttles FS, GS and H, and W.

day_typevarchar

weekday or weekend.

periodvarchar

peak or offpeak.

num_timepoints_passing_wait_assessmentbigint

Scheduled timepoints where the gap between trains met the standard.

num_sched_timepointsbigint

Scheduled timepoints.

wait_assessmentdouble

Share that met the standard, from 0 to 1.

  • To combine rows, add the counts, then divide, as for terminal on-time performance.
  • It looks at a sample of stops, not every station, and at scheduled gaps, not at riders. A line can score well while a station between the timepoints waits long.
  • Line names are not steady, as in terminal on-time performance: list each line with its first and last month before you add lines up.
  • A month appears once the MTA publishes it, usually in the next month.
  • Terms: NY open data. Credit the Metropolitan Transportation Authority. Source dataset: MTA Subway Wait Assessment.