VU BreakThru

Home » Data Science Methods for Smart City Applications » Safety Analytics using Motor Vehicle Accident Data

Safety Analytics using Motor Vehicle Accident Data

Posted by on Thursday, May 31, 2018 in Data Science Methods for Smart City Applications, News.

This spring, the University Course Data Sciences Methods for Smart City Applications was offered to Vanderbilt University undergraduate and graduate students for the first time. The course was designed to examine and address issues facing cities and metropolitan areas by bringing together concepts and methodologies from systems engineering, data sciences and machine learning, modeling and simulation, optimization and social sciences. Students reviewed these topics and focused on machine learning, Python programming and script generation, qualitative methods for data collection, and ethics issues related to data collection and privacy.

A significant portion of the semester was devoted to team projects with students working together. The projects covered four separate themes: transportation, energy, gentrification and transportation, and emergency response. Each team worked with a faculty mentor and a graduate student advisor to research their projects, implement qualitative and quantitative methods for data collection, and analyze the data to answer research questions. The students presented their findings at the end of the semester and submitted reports.

This is the third in a series of four student-written blogs describing the projects.

Safety Analytics using Motor Vehicle Accident Data
Written by Chris Morrison, Alexis Mundo, Hanna Nell and Andrew C. Standford

The Safety and Analytics group was tasked with constructing, modeling and ultimately optimizing resource allocation of Metro Nashville Fire Department (MNFD) emergency vehicles for motor vehicle accidents. The motivation for this project was based on shortening response times to incidents by using historical data of motor vehicle accidents. It gave us a chance to explore and practice mixing different kinds of research with computer programming in order to solve challenges that can help drive future innovation of smart cities. The course provided the opportunity to tackle the problem from a multidisciplinary perspective, emulating that of a real work environment.

We were able to visit fire stations and Nashville’s dispatch center in order to gauge how dispatch protocols work in conjunction with technology, with the ultimate goal of creating safer urban environments. Although we learned that their computer-aided dispatch system works extremely well, our analysis enabled us to find areas for improvement. For example, we concluded that future steps for stakeholders could include integrating incident prediction models with an analysis of different dispatch policies in order to assess efficiency and effectiveness of policies in place and future policies, allowing for better geographical allocation of emergency response resources. This could be a crucial next step, especially given our acquired knowledge of the scarcity of personnel and resources the MNFD have on hand.


Leave a Reply