Citation

Liu XC, Taylor J, Porter RJ, Wei R. J. Intell. Transp. Syst. 2018; 22(6): 530-546.

Copyright

(Copyright © 2018, Informa - Taylor and Francis Group)

DOI

10.1080/15472450.2018.1444484

PMID

unavailable

Abstract

The rapid expansion of bikeshare programs nationwide provides opportunities to gain insights on the optimal development of multimodal networks and bike-friendly environments. The profusion of trajectory-level data produced by bikeshare systems allows for information extraction on users' route preferences and, if modeled properly, will lead to a greater understanding of road characteristics that are appealing to bikeshare users. Leveraging Global Positioning System (GPS) data obtained from the GREENbike program, this study proposes a method to characterize roadways (e.g. collector, peripheral road, attractive road, and local road) on the basis of a variety of network centrality functions. The methodology is able to uncover the structure of the underlying transportation network and identify locations of critical bicycle infrastructures. A series of centrality measures, including degree, shortest-path betweenness, and random-walk betweenness centrality are implemented to determine the roadway classifications. Their suitability and usability for this purpose is then explored and discussed at length through a sensitivity analysis. The method can be applied to any bikeshare system that has access to trajectory-level (i.e. GPS, crowdsourcing) data for identifying road attributes that are appealing to bike users.

RESULTS can effectively guide future investment choices.

Language: en

Keywords

bikeshare; centrality; network modeling; road classification; trajectory