A tale of two stations: Analyzing metro ridership with big data

Abstract

This paper presents a multi-dimensional case study of the Beijing metro system. In particular, we examine two non-transfer stations, Zaoying and Jiangtai, which are on the same metro line in central Beijing. Multi-source and heterogeneous data are integrated to analyze and diagnose the drastically different metro ridership at the two stations. These include transit smart card data, taxi GPS data, network data, Point of Interest data, demographic data, online second-hand property price data, cell phone signalling data, and bike sharing data. The different utilization of metro system at these two locations is attributed to a number of factors pertaining to transportation infrastructure, built environment, demographic composition, commuting patterns, and connectivity of multi-modal transit networks. The findings suggest the importance of local accessibility of the metro stations as well as its connectivity with the rest of the transit system, in order to maximize the transport capability of the metro system. Our analysis also highlights the benefit of collecting and analyzing fine-granularity data in order to identify key bottlenecks and inefficiencies in the transportation system, as conventional macroscopic transportation planning data do not sufficiently capture the local accessibility and mobility in an urban environment.