Exploring factors affecting route choice of cyclists: A novel varying-contiguity spatially lagged exogenous modeling approach

Abstract

Cycling is one of the main transport modes and cycling infrastructure is strongly embedded in transport infrastructure in the Netherlands. Nonetheless, the bicycle network still undergoes frequent improvements and expansions. One of the critical elements in deciding on improvements and expansions is to understand the route choice of cyclists, which helps identify bottlenecks in bicycle flows and substantiate the need for new bicycle infrastructure. Yet, the factors affecting the route choice of cyclists are still not fully understood. To address this, we develop a varying-contiguity spatially lagged exogenous (VCSLX) model and analyze the probability of a cyclist choosing a certain segment based not only on the characteristics of that segment but also considering the characteristics of its neighbors along a route. Characteristics that are included in this study are the presence of bicycle infrastructure, traffic control installations and artificial lighting, as well as pavement type, bicycle and motorized-vehicle volumes and different land-use zones. The model involves the analysis of the observed routes extracted from cycling trajectories from Fietstelweek data, as well as corresponding hypothetical shortest path routes identified from the origin-destinations of the observed trips and the cycling network. The results of the study can help to understand the factors convincing cyclists to deviate from the shortest possible routes. The study contributes to the current literature by focusing on the underexplored aspect of spatial dependencies between route segments in the route choice of cyclists.