Jonathan Jay, JD MA
Harvard TH Chan School of Public Health
Thursday, September 20, 2018 9:00-10:00 AM CRB Austrian Auditorium
Environmental features, including tree cover and vacant lots, have been found to help explain the differences in gun violence risk from one city block to another. This project uses deep learning to extract information about the physical environment from high-resolution satellite imagery, then estimates shooting risk for every block in Philadelphia. Our findings challenge the dominant “hot spots” approach for predicting where victimization will occur – i.e., based on spatial proximity alone – and suggest that the environment may be particularly important for understanding the intersection of race and urban violence.
Jonathan Jay, DrPH JD, studies urban health risks using machine learning and spatial analytic methods. He is based at the Harvard T.H. Chan School of Public Health as a postdoctoral fellow for the Firearm-safety Among Children and Teens Consortium (FACTS), administered from the University of Michigan, and as a research fellow for the Computational Epidemiology Group at Boston Children’s Hospital. His long-term projects include predicting fires, overdoses and other emergency medical incidents for Portland (OR) Fire & Rescue. Originally trained as an attorney and bioethicist, Jon earned a doctorate in public health from Harvard in 2018.