Exploring the Potentials of Using Crowdsourced Waze Data in Traffic Management: Characteristics and Reliability

Real-time traffic information is essential to a variety of practical applications. To obtain traffic data, various traffic monitoring devices, such as loop detectors, infrastructure-mounted sensors, and cameras, have been installed on road networks. However, transportation agencies have sought alternative data sources to monitor traffic, due to the high installation and maintenance cost of conventional data collecting methods. Recently, crowdsourced traffic data has become available and is widely considered to have great potential in intelligent transportation systems. Waze is a crowdsourcing traffic application that enables users to share real-time traffic information. Waze data, including passively collected speed data and actively reported user reports, is valuable for traffic management but has not been explored or evaluated extensively. This dissertation evaluated and explored the potential of Waze data in traffic management from different perspectives.

First, this dissertation evaluated and explored Waze traffic speed to understand the characteristics and reliability of Waze traffic speed data. Second, a calibration-free incident detection algorithm with traffic speed data on freeways was proposed, and the results were compared with other commonly used algorithms. Third, a spatial and temporal quality analysis of Waze accident reports to better understand their quality and accuracy was performed. Last, the dissertation proposed a network-based clustering algorithm to identify secondary crashes with Waze user reports, and a case study was performed to demonstrate the applicability of our method and the potential of crowdsourced Waze user reports.