In Transportation
UNLABELLED : Determining bike-sharing usage patterns and their explanatory factors on demand is essential for the effective and efficient operation of bike-sharing systems (BSSs). Most BSSs provide different passes that vary with the period of use. However, studies investigating the differences in usage patterns are rare compared to studies conducted at the system level, even though explanatory factors depending on the type of pass may cause different characteristics in terms of usage patterns. This study explores the differences in the usage patterns of BSSs and the impact of explanatory factors on the demand depending on the type of pass. Various machine learning techniques, including clustering, regression, and classification, are used, in addition to basic statistical analysis. As observed, long-term season passes of over six months are mainly used for transportation (especially commuting), whereas one-day or short-term season passes seem to be used more for leisure than for other purposes. Furthermore, differences in the purpose of bike rentals seem to cause differences in usage patterns and variations in demand over time and space. This study improves ther understanding of the usage patterns that appear differently for each pass type, and provides insights into the efficient operation of BSSs in urban areas.
SUPPLEMENTARY INFORMATION : The online version contains supplementary material available at 10.1007/s11116-023-10371-7.
Kim Kyoungok
2023-Feb-21
Clustering, Demand prediction, Logistic mixed model, Public bicycle system, Spatial regression
