M.A. Farrag, A.Z. Elabdeen Heikal, M. Shawky Ahmed, A. Osama Amer
Traditional methods for investigating road safety mainly depend on crash data analysis, which is not always available, especially in developing countries. To overcome this shortage, new road safety tools/techniques have been introduced. This paper examined crowdsourcing data as a surrogate measure for accident data. This combined with users’ internet-based surveys can be analyzed to provide indications of road safety situation. Two major neighborhoods in Cairo, Egypt, were selected for the study. Two sets of data were considered: crowdsourced crash data and an online questionnaire. Data analysis revealed a strong correlation between the two datasets in the case of road segment safety identification and road crash causes. According to both sets of data, speeding and random pedestrian crossing are the primary causes of traffic accidents. Multinomial logistic regression models were devised to identify the variables that significantly influenced (1) exceeding the speed limit by drivers and (2) unsafe pedestrian road crossing behavior. Lastly, this paper has introduced validation of applying the introduced crowdsourced data for investigating road safety and road users’ behavior in Egypt and elsewhere when the traditional approach of using crash data is missing. Finally, policy recommendations are provided to improve drivers’ and pedestrians’ behaviors.
Keywords: crowdsourced data; road safety; internet-based survey; speeding; pedestrian crossing