While the first bicycle was invented over two hundred years ago, the two-wheeled mechanical wonder appears to be going through a renaissance. Compared to gas-guzzling cars, bikes are not only more environmentally friendly, but also offer benefits to human health. It’s no wonder then, that in land-constrained countries like Singapore, bike-share systems have become popular as an alternative to vehicles and a complement to public transportation systems.
Bike-share systems can be either station-based or dockless. Compared to station-based systems, dockless systems lack defined origin and destination points, making their use unpredictable and difficult to study. This is one reason why most research on bike-share usage thus far has focused on station-based systems, said Jie Song, a Research Scientist at A*STAR’s Institute of High Performance Computing (IHPC).
To fill this gap, Song and his colleagues developed an analytical approach to pinpoint when and where bike riders used dockless systems and applied them in Singapore. Their study forms part of a three-year ongoing project run by the Land Transport Authority of Singapore to simulate multiple modes of transport across the island using spatial autocorrelation and community modeling.
“The spatiotemporal analyses of shared-bike trips are capable of identifying ‘hotspot’ regions where high numbers of bike trips occur, which period of the day such hotspots form, and the average length of these bike trips,” Song noted.
By taking GPS data from all active bike-share providers in Singapore across eight continuous days—including weekdays, weekends and a public holiday—the researchers identified six clear hotspots around subway line interchanges: four in residential heartland towns like Yishun and Jurong East, and two downtown. This suggests that local riders use bikes to address the first-and-last mile problem, otherwise known as the initial and final leg of a trip to and from public transport hubs.
They also found larger “communities” on weekdays than weekends, indicating that Singaporeans may not simply use shared bikes for the first-and-last-mile problem; rather, end-to-end trips may be more common than expected. However, the researchers note that these interpretations are based on limited data and require validation.
Beyond Singapore, the team’s analytical approach can also be applied to systems elsewhere. “Given a dockless bike-share dataset in other major cities, the analyst can easily apply the proposed approach to generate a series of insights about the spatiotemporal characteristics of bike-share usage,” Song said.
Moving forward, their data-informed approach can be used to advise bike-share operators on the ideal location for bike-share facilities as well as urban planners on designing cycling-friendly towns, suggested Song and his co-authors.
The A*STAR-affiliated researchers contributing to this research are from the Institute of High Performance Computing (IHPC).
