Author ORCID Identifier
Year of Publication
Master of Science in Civil Engineering (MSCE)
Dr. Mei Chen
Traffic data obtained through crowdsourcing are becoming more accessible to traffic agencies due to advancements in smartphone technology. Traffic managers aim to use this data to complement their conventional sources of data and provide additional context in their analysis. In this study, Waze incident alerts are integrated with GPS-Probe speed data and Kentucky State Police (KSP) crashes to assess their impact on traffic flow and safety on freeways in Kentucky. The analysis showed that the presence of a vehicle on the shoulder is associated with about 36.7% of freeway crashes in Kentucky. The presence of a vehicle on the shoulder coupled with congestion were 11.7% of the crashes. As such, the correlation between vehicle on shoulder, congestion and crashes was significant. Albeit present within the vicinity of 7.4% of crashes, the presence of a vehicle in the travel lane did not show as having a significant correlation with crashes. Linking Waze crash alerts with crashes and assessing their spatiotemporal patterns, it is found that Waze crashes are spatially accurate and hence could be used as an alternate source for identifying crashes, sometimes earlier, in Kentucky and hence cutting down incident response and clearance times. The data used in this study and the analytical methods employed offer much needed insight into the potential of crowdsourced traffic incident data for traffic monitoring to ensure safety.
Digital Object Identifier (DOI)
Boasiako Antwi, Eugene, "ASSESSING FREEWAY CRASH RISK USING CROWDSOURCED WAZE INCIDENT ALERTS" (2021). Theses and Dissertations--Civil Engineering. 108.