Abstract

Spatially disaggregate Enumeration District (ED) level data for London is used in an analysis of various area-wide factors on road casualties. Data on 15335 EDs was input into a geographic information system (GIS) that contained data on road characteristics, public transport accessibility, information of nearest hospital location, car ownership and road casualties. Demographic data for each ED was also included. Various count data models e.g., negative binomial or zero-inflated Poisson and negative binomial models were used to analyze the associations between these factors with traffic fatalities, serious injuries and slight injuries. Different levels of spatial aggregation were also examined to determine if this affected interpretation of the results. Different pedestrian casualties were also examined. Results suggest that dissimilar count models may have to be adopted for modeling different types of accidents based on the dependent variable. Results also suggest that EDs with more roundabouts are safer than EDs with more junctions. More motorways are found to be related to fewer pedestrian casualties but higher traffic casualties. Number of households with no car seems to have more traffic casualties. Distance of the nearest hospital from EDs tends to have no significant effect on casualties. In all cases, it is found that EDs with more employees are associated with fewer casualties.