Modelling cyclist route choice in the Netherlands

In recent years, the bicycle has been gaining in popularity amongst governments, city planners, and residents for its potential to address many urban challenges, such as emission reduction, efficient use of public space, and public health (Kennisinstituut voor Mobiliteitsbeleid, 2020). As governments weigh investment options for cycling infrastructure, knowledge about cyclists' preferences as well as the ability to predict cyclists' travel patterns with a transportation model can offer important insights during the decision-making process. In the past 30 years, several stated preference surveys about cyclists' preferences have found that cyclists consider infrastructural and environmental factors as important when choosing their route. However, studies on route choice behavior have rarely been validated with observed data, and the small number of revealed preference studies to this date show mixed results depending on local context and methodology. In addition, knowledge about route preferences is seldom incorporated in transportation models to improve the realism of bicycle traffic predictions. This study addresses these gaps by estimating a discrete choice model for route choice based on observed GPS bicycle trips (revealed preferences) and infrastructural data, which is subsequently implemented in an existing transport model and validated as a more realistic method of predicting cyclists' route choices.