LAUSR

Abstract This paper explores effects of urban the built environment on traffic accidents in the context of a developing country. The basic objective of this study is to use simple and practical models to analyze traffic accidents, employing similar factors used by studies conducted in developed countries. The data collected at 107 road segments between the years 2008 and 2010 in Eskisehir (Turkey), include accidents with fatalities and injuries. Two sets of models have been developed: the first set uses binary logit models of traffic safety and the second set uses count data regression models to estimate occurrences of different accident types. Both sets consider three accident types: pedestrian-vehicle, vehicle-vehicle, and all accidents combined. Explanatory variables used in those models are derived from ambient land use characteristics, road segment properties, and traffic flow characteristics. Based on the results and our accompanying discussion, we outlined three types of direct policy implications. First, public transit should be regulated by relocating public transportation stations and by taking safety measures around stations. Second, traffic conditions and street networks should be improved by coordinating urban planning, street design, and transportation planning activities. Third, land use decisions should be reviewed focusing on taking precautions in areas associated with accident-increasing variables, rather than changing mixed land uses to monotonous land uses.