Seminar Details
Machine Learning and Ballistics: a Data Analytics Project in Partnership with the New Jersey State Police
- Start Date: November 15, 2018
- Event Start Time: 12:00 PM
- Event End Time: 1:00 PM
- Seminar Series: CCICADA Seminar Series in Homeland Security
- Presenter(s): Christie Nelson - DIMACS
- Event Location: DIMACS Conference Room | Rutgers University | CoRE Building, Room 433 | 96 Frelinghuysen Road
- Presentation Type: Stand Alone Presentation
- Abstract:
Christie Nelson and students in the Professional Masters program in Business and Science have partnered with the NJ State Police Forensics Lab on a ballistics data analysis project, utilizing machine learning (predictive and unsupervised models), data visualization, descriptive data analysis, and data cleaning techniques. We utilized data from a few different ballistics databases like NIBIN and Gun-ops, and examined data around when ballistics crimes were committed, their location, those that have a weapon used in multiple incidents, and many other questions of interest to the NJ State Police. Analysis included questions such as: what is the duration (time) between linked shooting events, where is the next crime likely to occur, what kinds of weapons are used most frequently in crimes (and are there variations between weapons used once vs those used repeatedly), and more.
This work was most recently cited in an Aug 2018 report done by the National Crime Gun Intelligence Governing Board.
It was also presented to the ATF, state and local law enforcement, a county prosecutor's office, the Police Foundation, the Center on Policing, and more (as described here), and was separately was presented to NJ Ballistics Community of Interest working group meeting (May 2018).