Data & Methods

This project compares two of the world’s busiest urban rail systems — the MTA’s Brooklyn-Manhattan Transit (BMT) group and JR East’s Yamanote Line — across four dimensions: network structure, passenger demand, service frequency, and reliability.

Data Sources

MTA (New York City)

MTA data is publicly available through the MTA Open Data Portal and covers 2024–2025.

GTFS — static schedule files (routes, trips, stop times, shapes) used to map the BMT network and compute scheduled trip frequency
Hourly ridership — average riders by hour of day, segmented by weekday and weekend
On-time performance (OTP) — monthly terminal OTP by day type, used as the primary reliability metric

JR East (Tokyo)

JR East data is more limited in public granularity than MTA. Three sources were combined:

GTFS via ODPT — Open Data Platform for Public Transportation, used for the Yamanote schedule, stop locations, and headway computation
Passenger survey (JSON) — annual station-level boarding counts published via ODPT
Annual station boardings — daily average boardings per station published at jreast.co.jp, covering FY2023 and FY2024

Key Limitations

No hourly ridership for Yamanote. JR East does not publish intraday ridership publicly. Frequency charts use scheduled departures as a supply-side proxy.
Different reliability definitions. MTA measures on-time performance at the terminal. Japan’s standard measures departure to within seconds of schedule — these are not directly comparable metrics.
Boarding counts only. JR East station data counts boardings (entries) only. MTA turnstile data counts both entries and exits separately.
Headways computed from scheduled GTFS, not observed operations. Actual headway variance during disruptions is not captured.

Tools

Analysis and visualization were built entirely in R using tidyverse, plotly, leaflet, scales, and sf. The site is built with Quarto.

For the complete methodology, analytical decisions, and limitations discussion, see the full white paper.