Towards spatiotemporal integration of bus transit with data-driven approaches

TL;DR

This work tackles the problem of enabling seamless bus-line transfers by integrating spatiotemporal data without relying on fixed timetables. It introduces two data-driven algorithms: an itinerary-detection method that aligns GPS logs with line itineraries through map matching, temporal sequencing, and interpolation, and a bus-stop clustering method that defines virtual terminals within a walking radius to support single-fare transfers. The approach yields a substantial increase in itinerary traceability from $68.83 ext{%}$ to $99.33 ext{%}$ and can reduce total travel distances from $22.4$ km to $12.3$ km, at the cost of increased transfers ($\sim 2 \to 4$). Overall, the methods demonstrate meaningful potential to boost transit efficiency and attract riders, while highlighting needs for improved travel-time estimation and transfer timing models for real-world deployment.

Abstract

This study aims to propose an approach for spatiotemporal integration of bus transit, which enables users to change bus lines by paying a single fare. This could increase bus transit efficiency and, consequently, help to make this mode of transportation more attractive. Usually, this strategy is allowed for a few hours in a non-restricted area; thus, certain walking distance areas behave like "virtual terminals." For that, two data-driven algorithms are proposed in this work. First, a new algorithm for detecting itineraries based on bus GPS data and the bus stop location. The proposed algorithm's results show that 90% of the database detected valid itineraries by excluding invalid markings and adding times at missing bus stops through temporal interpolation. Second, this study proposes a bus stop clustering algorithm to define suitable areas for these virtual terminals where it would be possible to make bus transfers outside the physical terminals. Using real-world origin-destination trips, the bus network, including clusters, can reduce traveled distances by up to 50%, making twice as many connections on average.

Figures (16)

• Figure 1: Example of an itinerary $iti=(p_1, p_2, p_3, p_4, p_5)$ in full line and a $log=((l_1,t_1), (l_2,t_2), (l_3,t_3), ( l_4,t_4), (l_5,t_5))$ in dashed line.
• Figure 2: Programmed itinerary of line 829 that starts at point 1, passes through intermediate points 2 to 10, and returns to the starting point.
• Figure 3: Region of uncertainty in which the map matching algorithm generates a marking error.
• Figure 4: Interpolation error in seconds on line 829 for different values of $w$ (a missing midpoint corresponds to $w=2$).
• Figure 5: Line type interpolation error.