How Could Space Syntax Be Applied in Traffic Accident Frequency Modelling-An Example Based on Chinese City

The Traffic Accidents (TAs) study plays a significant role in transport strategy making and planning. This study incorporates traditional parameters, such as geographical, environmental, traffic factors and Space Syntax (SS) variables into accident modelling to better understanding causal mechanisms. The Negative Binomial models are applied to accident modelling based on the accident dataset in a Chinese city, Nanning. We assess the effect of SS variables on TAs frequency, controlling for geometric and land-use factors. Results show that SS metrics significantly improve model performance across all accident types. Higher angular connectivity and proximity to the old city centre could increase TA frequency. These findings examine how the spatial configuration of the network impacts TAs and the reliability and safety perspective of the road network.