2024/02/07 by Berrill, Peter, Nachtigall, Florian, Javaid, Aneeque +3
#600 Technik #Klimawandel #Maschinelles Lernen #Medizin #Mobilität #Stadtform #Verkehrsmittelwahl #angewandte Wissenschaften::620 Ingenieurwissenschaften::629 Andere Fachrichtungen der Ingenieurwissenschaften #climate change #machine learning #mobility #mode choice #urban form
paper · doi:10.14279/depositonce-19831
Steady growth in global greenhouse gas emissions from transport is driven by growing demand for car travel. A sizable body of research investigates influences of urban form on travel behavior, but few European studies illustrate variation of these influences across multiple cities and countries using disaggregated data. Here, we compare car ownership and mobility patterns, and we use gradient boosting decision tree and regression models to investigate urban form influences on travel distances, vehicle ownership, and mode choice across nineteen diverse European cities. Residential proximity to the city center is the urban form feature with greatest predictive importance for trip distances, car ownership and car mode choice. The exponential reduction of car use with higher population density is clearly demonstrated with aggregate city data. We detect nonlinear relationships between urban form and modelled outcomes, identify urban form thresholds for sustainable mobility, and suggest targeted policy interventions.