This workshop aims to explore the role of open data in cycling planning and its integration into agent-based models, focusing on the Berlin Metropolitan Area. Previous research has underlined the impact of unsafety concerns, created by still car-dominated urban environments, on bicycle usage and behavior. While various models have incorporated safety perceptions, their subjective nature introduces complexity that might not be necessary. In the workshop, we will also explore how modeling should guide cycling planning and how immediate infrastructure interventions should be prioritized in reversing above negative concerns. Furthermore, we want to examine how different data sets can be used in cycling planning.
The first part of the workshop will focus on conducting meaningful qualitative comparisons considering different dimensions of cycling behavior. Berlin presents an ideal case study due to its availability of datasets. These are: a) reported incidents or accidents and rides, b) road infrastructure data with estimated perceived safety rates per road link, and c) bicycle traffic counts. After presenting some trends, it will be examined whether actual safety metrics sufficiently explain cycling behavior or if perceived safety plays a more significant role by comparing the findings from steps a), b) and c) to findings of existing studies. Furthermore, the relationship between cycling infrastructure and observed mobility patterns will be discussed in the workshop.
Next, findings from agent-based simulation experiments using MATSim (Multi Agent Transport Simulation) will be presented. The discussion will center on identifying deviations between modeled and real-world cycling behaviors. The importance of these outputs in the planning process will be critically assessed, considering whether recalibration of existing models is necessary or if more advanced stochastic simulation approaches should be developed.
These discussions with the participants will contribute to defining the relationships between data, modeling, and policy decisions to answer how research can support a data-driven yet pragmatic approach to cycling planning.

This abstract was prepared by Gregor Rybczak, Simon Meinhardt and Panagiotis G. Tzouras