No abstract

No abstract

The spatial nature of crash data highlights the importance of employing Geographical Information Systems (GIS) in different fields of safety research. Recently, numerous studies have been carried out in safety analysis to investigate the relationships between crashes and related factors. Trip generation as a function of land use, socio-economic, and demographic characteristics might be appropriate variables along with network characteristics and traffic volume to develop safety models. Generalized Linear Models (GLMs) describe the relationships between crashes and the explanatory variables by estimating the global and fixed coefficients. Since crash occurrences are almost certainly influenced by many spatial factors; the main objective of this study is to employ Geographically Weighted Poisson Regression (GWPR) on 253 traffic analysis zones (TAZs) in Mashhad, Iran, using traffic volume, network characteristics and trip generation variables to investigate the aspects of relationships which do not emerge when using conventional global specifications. GWPR showed an improvement in model performance as indicated by goodness-of-fit criteria. The results also indicated the non-stationary state in the relationships between the number of crashes and all independent variables.

No abstract

Two-way two-lane roads constitute a major proportion of rural road networks in Iran. The work described in this paper aimed to investigate the factors that influence overtaking manoeuvres using traffic conflict techniques to collect field data related to dangerous overtaking. The process of data collection was conducted by means of video recordings and other heuristic methods using inductive loop detectors. As one of the most important underlying independent variables, the percentage of time spent following was simulated. The classification and regression tree method (a common data mining approach) was used to identify important variables and explore the cut points of continuous variables. No-passing distance, road width, mean speed and vehicle time spent following were identified as the most important variables, and the results indicated a direct relationship between no-passing distance and the number of dangerous overtaking manoeuvres. Road widths less than the standard values and an increase in mean speed were also found to lead to an increase in the number of dangerous overtakings.