nyc-transit-data/turnstile_daily_counts_2021

NYC Subway Turnstile Counts Data aggregated by day and station complex for the year 2021. Updated weekly, usually on Monday mornings. (The pipeline is mostly automated, but must be manually triggered)

Where the Data Came From

This aggregation was created from weekly raw turnstile counts published by the New York MTA at http://web.mta.info/developers/turnstile.html

The raw data were imported into a postgresql database for processing, and aggregated to calendar days for each station complex.

The process is outlined in this blog post, and the code for the data pipeline is available on github.

Caveats

This aggregation is a best-effort to make a clean and usable dataset of station-level counts. There were some assumptions and important decisions made to arrive at the finished product.

• The dataset excludes turnstile observation windows (4 hours) that resulted in entries or exits of over 10,000. This threshold excludes the obviously spurious numbers that come from the counters rolling over, but could include false readings that are within the threshold.

• The turnstile counts were aggregated to calendar day using the timestamp of the end of the 4-hour observation window + 2 hours. An observation window that ends at 2am would count for the same day, but a window ending between midnight and 1:59am would count for the previous day.

• The last date in the dataset contains a small number of entries and exits that will be aggregated into the next week’s worth of data, and should not be used.

PATH and Roosevelt Island Tramway

The dataset also includes turnstile counts for the PATH train system and the Roosevelt Island Tramway