Where are public bikes? The decline of dockless bike-sharing supply in Singapore and its resulting impact on ridership activities. (April 2021)

• Record Type: Journal Article
• Authors: Song, Jie; Zhang, Liye; Qin, Zheng; Ramli, Muhamad Azfar
• Is Part Of: Transportation research. Volume 146(2021)
• Page Start: 72
• Page End: 90
• Publication Date: 2021-04
• Subjects: Shared bicycle -- Clustering -- Spatiotemporal -- Network analysis; Transportation -- Research -- Periodicals; 388.011
• DOI: 10.1016/j.tra.2021.01.016
• Languages: English
• ISSNs: 0965-8564

Abstract

Highlights: A before-and-after experiment was performed to examine the association of bike sharing services and cycling activities. The impact of reduced bike sharing services on riders' behavior was analyzed. Declined bike supply impaired the diverse landscape of spatial communities of cycling flows. Less public bikes may lead to the increased inaccessibility of different population groups to bike sharing facilities in certain areas. Abstract: Active mobility modes have become a significant alternative mode of transport within all major urban cities. This could be partly attributed to the rise and fall of dockless bike-sharing services in the last three years which significantly changed the transport ecosystem in these cities. However, the various complex relationships between the supply of bike-sharing services, the generated demand as well as the cycling dynamics and behavior of subscribers is not well understood. In this article, we examine how the changes in bike supply impacted the cycling activities of a dockless bike-sharing system in Singapore. We employed a clustering algorithm and modelled the network community detection tool into a three-month cycling GPS dataset and compared the spatiotemporal pattern and network structures of biking journeys for two major time periods (1) January 2019 and (2) March 2019 between which a significant number of bike-sharing companies exited the industry. This resulted in a drastic drop in the number of available shared bikes for the latter period. Our results show that prior to the decline of the bike supply, we identified three types of bike journeys that shared similar temporal patterns, indicating different riders, namely casual, moderate, or enthusiastic users. Over 90 percentage of enthusiastic riders disappeared after the supply of public bikes decreased. Our network analysis further showed that nine out of ten observed communities either contracted or collapsed after the supply downsize. We therefore hope that urban and transportation planners and scientists would be able to use these findings to further understand the dynamics of supply demand of bike-sharing systems and help improve planning considerations in the future.