Degree Type


Date of Award


Degree Name

Doctor of Philosophy


Industrial and Manufacturing Systems Engineering


Industrial and Manufacturing Systems Engineering

First Advisor

Stephen B. Gilbert

Second Advisor

Anuj Sharma


The widespread availability of internet and mobile devices has made crowdsourced reports a considerable source of information in many domains. Traffic managers, among others, have started using crowdsourced traffic incident reports (CSTIRs) to complement their existing sources of traffic monitoring. One of the prominent providers of CSTIRs is Waze. In this dissertation, first a quantitative analysis was conducted to evaluate Waze data in comparison to the existing sources of Iowa Department of Transportation. The potential added coverage that Waze can provide was also estimated.

Redundant CSTIRs of the same incident were found to be one of the main challenges of Waze and CSTIRs in general. To leverage the value of the redundant reports and address this challenge, a state-of-the-art cluster analysis was implemented to reduce the redundancies, while providing further information about the incident. The clustered CSTIRs indicate the area impacted by an incident and provide a basis for estimating the reliability of the cluster. Furthermore, the challenges with clustering CSTIRs were described and recommendations were made for parameter tuning and cluster validation.

Finally, an open-source software package was offered to implement the clustering method in near real-time. This software downloads and parses the raw data, implements clustering, tracks clusters, assigns a reliability score to clusters, and provides a RESTful API for information dissemination portals and web pages to use the data for multiple applications within the DOT and for the general public.

With emerging technologies such as connected vehicles and vehicle-to-infrastructure (V2I) communication, CSTIRs and similar type of data are expected to grow. The findings and recommendations in this work, although implemented on Waze data, will be beneficial to the analysis of these emerging sources of data.

Copyright Owner

Mostafa Amin-Naseri



File Format


File Size

97 pages