Department of Mathematics Presentation Announcement
Presenter: George Mohler, Boston College
Title: “Using data science to understand patterns of social harm”
Date: February 13, 2025
Time: 12:30 -1:45 PM
Room: YR205
In this talk we give three examples of how statistical modeling and machine learning can be applied to social harm data. We first investigate the relationship between law enforcement drug seizures and subsequent opioid overdoses. Using two-sample randomization tests, we show that seizures lead to an increased risk of overdoses nearby in space and time. We then discuss implications of this effect and potential interventions that might be conducted. In the second example, we use a statistical model to understand the recent “Kia Challenge” on TikTok. We then use the model, combined with information on past car theft waves, to forecast Kia and Sonata thefts in Los Angeles out to 2050. In the last example, we discuss a project on building low-cost gunshot detectors for community violence interruption workers in Denver. Using audio transformer neural networks on a Raspberry Pi, we are able to produce accurate devices for under $200. The devices are currently being tested for their efficacy in violence interruption efforts.
For more information, please contact the Department of Mathematics
Categorised in: General
This post was written by Charles, Amanda G.