Award Date


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Civil and Environmental Engineering and Construction

First Committee Member

Jee Woong Park

Second Committee Member

Mohamed Kaseko

Third Committee Member

Haroon Stephen

Fourth Committee Member

Pramen P. Shrestha

Fifth Committee Member

Ashok Singh


The fatalities, injuries, and property damage that result from traffic crashes impose a significant burden on society. Current research and practice in traffic safety rely on analysis of quantitative data from crash reports to understand crash severity contributors and develop countermeasures. Despite advances from this effort, quantitative crash data suffers from drawbacks, such as the limited ability to capture all the information relevant to the crashes and the potential errors introduced during data collection. Crash narratives can help address these limitations, as they contain detailed descriptions of the context and sequence of events of the crash. However, the unstructured nature of text data within narratives has challenged exploration of crash narratives. In response, this dissertation aims to develop an analysis framework and methods to enable the extraction of insights from crash narratives and thus improve our level of understanding of traffic crashes to a new level. The methodological development of this dissertation is split into three objectives. The first objective is to devise an approach for extraction of severity contributing insights from crash narratives by investigating interpretable machine learning and text mining techniques. The second objective is to enable an enhanced identification of crash severity contributors in the form of meaningful phrases by integrating recent advancements in Natural Language Processing (NLP). The third objective is to develop an approach for semantic search of information of interest in crash narratives. The obtained results indicate that the developed approaches enable the extraction of valuable insights from crash narratives to 1) uncover factors that quantitative may not reveal, 2) confirm results from classic statistical analysis on crash data, and 3) fix inconsistencies in quantitative data. The outcomes of this dissertation add substantial value to traffic safety, as the developed approaches allow analysts to exploit the rich information in crash narratives for a more comprehensive and accurate diagnosis of traffic crashes.


crash narratives; crash severity; natural language processing; semantic search; text analysis; text mining


Artificial Intelligence and Robotics | Civil Engineering | Computer Engineering

File Format


File Size

3600 KB

Degree Grantor

University of Nevada, Las Vegas




IN COPYRIGHT. For more information about this rights statement, please visit