Mining traffic data for the development of an accident warning application for tourists

Abstract

Tourist drivers belong to a category of drivers that are more vulnerable to road accidents due to their unfamiliarity of the road network at a destination. This paper presents a method followed to develop a tool that alert tourist drivers of their accident risks based on situational factors obtained from mobile phone sensors and knowledge distilled from historical records of traffic accidents. The knowledge necessary for the development of a context aware mobile accident warning application was extracted from a spatiotemporal analysis of historical accidents data, to identify patterns of conditions that lead to accidents. Results from this analysis were used to develop heuristics rules that were programmed in a mobile application. The developed system warns travelers of possible threats on the road network of Nicosia, given driver’s location and situational factors. The system aims to improve tourists’ safety.