nyc-subway-isochrones

Scripts (nodejs and R) for generating travel time isochrones from stations in the New York City subway.

Workflow

1. Start with the NYC Subway static GTFS file

data/google_transit.zip (unzipped to data/google_transit)

1. Run Rscript caclulate-station-times.R, which outputs a file {stop_id}.csv for each stop in the GTFS data. Each output file contains the travel times to each other station in the system for a given stop_id.

2. Run node trip-times-to-isocrhone.js, which converts each of the travel times CSVs generated in step 2 into a geojson FeatureCollection containing 10, 20, 30, and 40 minute isocrones.

Isochrone Creation Algorithm

1. For each station in the system, parse {stop_id}.csv and group all of the other stations into 10,20,30,40 minutes of travel time. All stops whose travel time is greater than 40 minutes from the origin are ignored.

2. For each travel-time group:

3. Determine the leftover amount of time to reach the maximum for that group. (10 minutes group - 7 minutes travel time = 3 minutes leftover) This is the amount of "walking" that can be done after leaving the train.

4. Translate the walking time by a fixed distance (e.g. 1.2 meters/second)

5. Generate a buffer for each station based on the walking distance.

6. Buffer the previous travel time group by a full 10-minute walk.

7. Dissolve the station buffers from #2 with the previous travel time group walking buffer from #3

8. Clip to the shoreline of NYC (data/nyc-shoreline-clip-multi.json)

9. Save as a geojson FeatureCollection in isochrones/{stop_id}.geojson

Other scripts

stops-to-geojson - converts stops.txt into a geojson FeatureCollection, or use in mapping.