Skip to main content

CRASH Predictive Analytics

Primary Investigators:
Hiba Baroud, Abhishek Dubey, Dan Work
Brief Description of Project:
The goal of this project is to build and deploy predictive analytics of highway incidents and apply it towards safety patrol vehicles deployment. The objectives of this research are to (i) identify the best practices for data storage, integration, and maintenance infrastructure for predictive modeling, (ii) develop state-of-the-art machine learning algorithms for predicting the risk of highway incidents, and (iii) collaborate with TDOT and THP to identify best practices for model integration with existing programs.

The undergraduate research assistant is expected to help with the implementation of the predictive analytics tool and assess the predictive accuracy for different types of models. Other research tasks will include data collection and processing as well as feature selection for identifying the most significant factors impacting the likelihood of a highway incident. In addition, they will be required to integrate the predictive model into a visualization tool.

Desired Qualifications:
The candidate is expected to have skills in coding and machine learning algorithms. Basic knowledge of statistical modeling and good written and communication skills are required.
Nature of Supervision:
The researcher will work directly with a graduate research assistant and will be supervised by the PIs on the project.
A Brief Research Plan (period is for 10 weeks):
Week 1-2 Literature review on machine learning models for crash prediction
Week 3-5 Implementation of predictive analytics tool
Week 6-7 Integration of predictive model into a visualization tool
Week 8-10 Preparing for presentation and paper
Number of Open Slots: 1
Contact Information:
Name: Hiba Baroud
Title: Assistant Professor
Department: Civil and Environmental Engineering