Investigating street accident characteristics and optimal safe route recommendation: A case study of New York City

• Published in: 2017 25th International Conference on Geoinformatics pp. 1 - 7
• Main Authors: Enbo Zhou, Shanjun Mao, Mei Li
• Format: Conference Proceeding
• Language: English
• Published: IEEE 30-10-2017
• Subjects: Accident Prone Street Detection; Accidents; Data processing; Motor Vehicle Collisions; Optimal Route Selection; Safety; Street Accident Characteristics
• ISSN: 2161-0258
• DOI: 10.1109/GEOINFORMATICS.2017.8090942

With the growing traffic accidents, cities meet the problem of traffic safety. In order to reduce the traffic accidents, vehicle collision's spatio-temporal distributions and characteristics need to be analyzed. This paper focuses on investigating the spatio-temporal distributions and characteristics of vehicle collisions and detecting the accident prone streets. Besides, an optimal safe route method is proposed based on the dangerous index defined in this paper. We calculated the number of collisions on each street by hour from 0 to 23 and got a collision-curve for each street to delineate the temporal characteristics of collisions which indicated the inherent functions and locations of each street. The streets were clustered into several types based on the collision-time curves to find different spatial patterns. A dangerous index was defined and calculated for each street. The accident prone streets were detected using the index to reveal the spatial distributions of dangerous streets. Finally, the optimal safe route method was tested on the data of New York City. This paper's methods and findings may contribute to the governance and planning of the city and the optimal safe routing method provides a new insight to route under some circumstances such as the school bus's travel.