Various international research projects have already revealed that the traffic behavior of cyclists is significantly influenced by the age and gender of cyclists as well as trip purpose. In addition, several studies showed that bicycle route choice depends largely on infrastructural factors (e.g. type of cycling infrastructure or surface conditions), environment (e.g. slope or scenery) and operational factors (e.g. maximum speed of motorized traffic or waiting times at intersections). Studies concerning the German context were able to prove these correlations, especially for the city of Dresden (Germany).
However, the results are only valid for the investigation area of the individual studies. It can be assumed that cycling behavior and route choice are not independent from local conditions (e.g. depending on the city size or topography). Therefore, a Germany-wide analysis of cycling behavior and route choice is required in order to not only make statements about German cyclists but also to investigate if and to what extent cycling behavior depends on specific influencing factors that could be identified in other countries as well. Knowledge of the relevant and city dependent influencing factors can provide useful information for transportation planning to promote cycling in cities with different characteristics.
The contribution presents first results of the research project ’Cycling behavior in Germany (RiD) - A multi-criteria approach to quantify local differences in cycling behavior and route choice’ funded by the German Research Foundation (DFG). The study is based on a Germany-wide GPS dataset on cycling, which includes more than 7 million bicycle trips from over 2,400 German cities (year of survey: 2022). In order to analyze cycling behavior, city categories (e.g. according to city size, topography) are defined. Bicycle trips are enriched with secondary data (e.g. infrastructure data) so that cycling and route characteristics can be determined. Subsequently, comparative analyzes of cycling behavior and route choice are carried out to derive insights into the influence of city characteristics on cycling behavior. For this purpose, classification and regression models using both statistics and machine learning methods are applied to the entire data set.
First results of the comparison reveal that cycling behavior differs between cities that show different city characteristics. Bicycle route choice, for instance, shows similar tendencies in different cities (e.g. existence of cycling infrastructure and a good surface positively influences route choice, whereas increasing distance or high slopes negatively influence route choice probability). However, the strength of the influencing factors differs from city to city. Cyclists react similarly, for instance, to increasing distance or maximum slopes along their route whereas they show differing reaction to decreasing surface quality or the existence of cycling infrastructure. This difference seems to depend on city characteristics such as city size as well as quality and quantity of existing infrastructure itself. Further results will be presented in the final contribution. The knowledge about differences in route choice and cycling behavior can be used to adapt network planning and improve traffic flow (e.g. to implement bicycle priority routes or green waves).