AI Analyzer for Revealing Insights of Traffic Crashes
The Innovations Deserving Exploratory Analysis (IDEA) product developed in this project is an open-source software tool that transportation engineers can use to analyze traffic crash data using Artificial Intelligence (AI), specifically Natural Language Processing (NLP). NLP has made great strides in understanding human language, as demonstrated by its success in tasks like sentiment analysis, intent detection, and question answering. By leveraging NLP, the developed tool can analyze traffic crash narratives and identify the factors that contribute to the severity of crashes. This provides a promising tool for extracting insights from traffic crash narratives, which allow us to gain a better understanding of traffic crashes and thus design more effective countermeasures. Traditionally, traffic safety analysis focuses mostly on the severity of traffic crashes using quantitative data available in crash reports, while the rich information in textual descriptions of crashes is largely underutilized due to our limited ability to process large amounts of data; therefore, very little is known about this information. The developed tool offers unprecedented opportunities to extract valuable information from a data source that has been vastly underutilized in the context of traffic safety. By assisting analysts in processing large amounts of crash narratives, the developed tool offers insights that other data types may not reveal, providing further opportunities to identify the crash factors that require urgent attention in the implementation of countermeasures to mitigate future traffic crashes.
- Record URL:
-
Supplemental Notes:
- This Innovations Deserving Exploratory Analysis (IDEA) project was conducted at the University of Nevada, Las Vegas.
-
Authors:
- Park, Jee Woong
- Arteaga, Cristian
- Publication Date: 2023-6
Language
- English
Media Info
- Media Type: Digital/other
- Edition: Final Report
- Features: Appendices; Figures; References; Tables;
- Pagination: 47p
-
Serial:
- NCHRP-IDEA Program Project Final Report
- Issue Number: 231
- Publisher: Transportation Research Board
Subject/Index Terms
- TRT Terms: Artificial intelligence; Data analysis; Innovation; Traffic crashes
- Identifier Terms: Innovations Deserving Exploratory Analysis (IDEA) (Program)
- Subject Areas: Data and Information Technology; Highways; Safety and Human Factors;
Filing Info
- Accession Number: 01889907
- Record Type: Publication
- Report/Paper Numbers: NCHRP IDEA Project 231
- Files: TRIS, TRB, ATRI
- Created Date: Aug 12 2023 2:21PM