Spatial Regression Analysis Using Poisson Regression: Applications in Studying Traffic Accidents

The study of spatial data and the application to the Poisson regression model, with special interest for the study of traffic accidents, is the objective of this work. Traffic accidents are a pressing problem with wide impacts on urban and road safety. As a result, knowing the pattern of these accidents is essential to urban planning and the development of effective safety measures. This study focuses on the computation of the Poisson regression model and the interpretation of traffic accident data as a dependent variable. It is assumed in this model that accidents happen at random but not independently and they depend on factors (explanatory variables) such as type of road, driver characteristics, character of road vehicle pedestrian other, etc., related to the probability that an accident occurs. Because of its characteristics for spatial data with count dependent variables, Poisson regression has the capability to establish a good model to predict the number of accidents under specific local conditions. The publication reports real-life applications, using traffic accident data collected in certain regions, for example, the spatial analysis which was performed to identify traffic accident hotspots for the 15 governorates of Iraq for the year 2023, and recommendations to improve road safety. The study also shows that the application of Poisson analysis may be an effective approach when strategic decisions are to be made in order to decrease accidents and improve security on the road. Finally, the study demonstrates that researchers should adopt state-of-the-art statistical tools when analyzing spatial information (e.g., Poisson regression), in the service of better practice and scientific solutions aimed at cutting traffic risk at high levels.