AI Analyzer for Revealing Insights of Traffic Crashes

A new AI-based tool has been developed and designed as a complement to classic statistical analysis in reaching a more comprehensive and accurate understanding of the factors that contribute to traffic crashes. It has the potential to help transportation agencies identify high-risk areas and develop targeted countermeasures to reduce the frequency and severity of crashes.

NCHRP IDEA Final Report for Project 231: AI Analyzer for Revealing Insights of Traffic Crashes, from TRB's IDEA (Innovations Deserving Exploratory Analysis) Program, details how the tool, which uses natural language processing to extract insights from textual crash narratives, has been shown to be able to successfully recover a large proportion of correlations in a dataset. It is also open-source, which means that other organizations can download and use it to analyze their own crash narratives.