Authors: Benjamin Acker*, University of Texas at Dallas, May Yuan, University of Texas at Dallas
Topics: Spatial Analysis & Modeling, Geographic Information Science and Systems, Transportation Geography
Keywords: network analysis, GIS, traffic accidents, risk, spatiotemporal, space syntax
Session Type: Paper
Start / End Time: 3:20 PM / 5:00 PM
Room: Bayside C, Sheraton, 4th Floor
Presentation File: No File Uploaded
This research aims to develop a model to assess the risk of a traffic accident occurring on any given segment of an urban road network using dynamic and static data. Several years of traffic accident data from the Dallas area are used to conduct network kernel density estimation of the historical risk of each road segment in the City of Dallas. Spatial risk measures will be determined based on the classification of accident densities on the street network. Next, we explore independent variables with spatial correlations to the spatial risk measures to build a model for spatial risk assessment. Space syntax is used to characterize street structure in Dallas and serve as a proxy for traffic flow and visibility in addition to other topological characteristics. Other independent variables include road characteristics (e.g. width) and dynamic situations (e.g. severe weather, time of day, day of the week, and nearby accidents). The influence of a nearby accident on the risk of a road segment, i.e. a near repeat, is identified using spatiotemporal network clustering. The proposed method would be useful to help identify risk street segments in space and time and identify spatial correlates for risk mitigation. Specifically, this spatial risk modeling is part of the SafeNET project funded by the National Institution of Standards and Technology (NIST) and in conjunction with the Dallas Fire-Rescue Department to improve the safety and response time of emergency vehicle dispatching.