15 minutes from City Hall: walk vs bike vs drive
One origin, three modes, one clock. From NYC City Hall, 15 minutes of walking reaches 2.75 km2, 15 minutes of cycling reaches 15.95 km2, and 15 minutes of freeflow driving reaches 151.62 km2: a 55x spread from the same corner of Manhattan. Bands at 5, 10, and 15 minutes per mode, computed on NYC's official LION street network. The three modes nest cleanly (containment 1.0 at every time cut: walk within bike within drive), so the map reads as three families of contours radiating from one point. Each band carries area_km2 and node_count in its feature properties. Companion to the five-borough walking map, which holds mode constant and varies the origin; this one holds the origin constant and varies the mode. Drive times use freeflow edge costs (no traffic adjustment) so results are reproducible.
Detrás del trabajo
Problema
How much does mode choice change reach from a single point? Reachability comparisons are usually eyeballed or produced with proprietary APIs; this uses only open city data and open-source routing.
Rol
I built the routing stack this analysis runs on.
Fuentes de datos
- NYC DCP LION street network (25d)
- NYC Geosupport for address search
- Traffic sources (NYC DOT volumes, TRANSCOM speeds) exist in the stack but were disabled for this run
Sistema de coordenadas
EPSG:4326 (WGS 84)
Cadena de herramientas
Notas del proceso
Graph built from LION with per-mode edge costs, one-way restrictions, and grade-separated crossings; drive uses freeflow speeds so results are reproducible. Reachable nodes computed with pgr_drivingDistance (Dijkstra) at 5, 10, and 15 minute cost limits per mode, then polygonized with ST_ConcaveHull and post-processed to cumulative unions per mode (the 10 minute band includes the 5 minute band). Bands are published as cumulative isochrones and rendered as outlines, so each contour is the true reachable-within-N boundary and the three modes stay readable where they overlap. City Hall is included as a point feature. Concave hulls can slightly over-include narrow water crossings at band edges.
Validación
Within-mode band nesting (5 within 10 within 15) checked in PostGIS after the cumulative-union step: intersection coverage = 1.0 for all 6 adjacent-band pairs (3 modes x the 5-10 and 10-15 transitions). Cross-mode containment at 15 minutes also measures 1.0 (walk within bike within drive). Per-band reachable node counts and areas are included in feature properties.
Licencia de datos
LION is NYC Open Data (free and open)