How flexible transportation services work in reality?- some insights from real-world observations

Public transportation in Japan currently faces serious challenges, including depopulation, dispersed low demand, and a shortage of drivers. To address these issues and cover wider areas with fewer drivers, flexible transport systems like demand-responsive transport (DRT) services are becoming increasingly popular, particularly in rural areas, thanks to recent advancements in Information and Communication Technologies (ICT). Despite the potential for reduced operational costs through more efficient service provision, overall costs often remain high due to increased operator and system costs. Improving the efficiency of services is crucial even for flexible transport systems. Understanding detailed traveler behaviors within these systems is essential for this purpose. Flexible transport systems often incorporate online booking and vehicle assignment systems, allowing for the automatic collection of booking data. By analyzing this data, we can gain insights into the behaviors of travelers and the patterns of bus stop utilization. This study utilizes booking data to examine the interactions between passengers and bus stops in flexible transport systems, with a particular focus on understanding and discussing patterns of regularity and variability in both traveler behavior and bus stop usage. The study uses nine years of booking data (2015–2023) from a mid-sized city in Gifu Prefecture, encompassing 845 passengers and 142,638 records. The analysis first explores the regularity of traveler behaviors and bus stop usage patterns, followed by a discussion on the flexibility or variability of vehicle movements. The results show that vehicle movements are primarily driven by regular high-frequency travelers, who use the service for commuting and returning home. This dominance often excludes low-frequency random travelers from accessing the service. Additionally, it is suggested that minimizing total operational costs may not adequately assign travelers onto vehicles, and the implementation of monthly passes may further reinforce the dominance of high-frequency travelers. These insights underscore the importance of considering service designs from various dimensions, including user behavior, spatial factors, and temporal patterns, for the effective optimization of flexible transport systems. • This paper analyzes nine years of booking data to examine travel patterns in an on-demand bus system. • Both passengers and bus stops were clustered, revealing insights into their relationship and usage patterns. • Frequent passengers dominate the on-demand service, especially during peak hours, limiting access for others. • Vehicle trajectories are concentrated during peak times but more dispersed during off-peak periods. • Recommendations include adjusting service areas, booking systems and pricing strategies to improve efficiency and fairness.