Towards sensing urban-scale COVID-19 policy compliance in new york city

Big data on the urban scale can enable many applications for improving city life and provide a more holistic understanding of urban life to researchers. While there are approaches to sense and model urban occupant behaviors using sound, radio frequency, and vision, how such behaviors are altered due to city governance and policies in response to emergencies such as a natural disaster or a public health crisis has been less explored. In this paper, we present a computer vision-based approach to capture patterns and interference in the urban life of New York City dwellers from March 2020 to August 2020. Using ~1 million images gathered with cameras mounted on ride-sharing vehicles throughout the city, we approximated the social proximity of pedestrians to understand policy compliance on the street. Our analysis reveals a correlation between policy violation and virus transmission. We believe that such big data-driven city-scale citizen modeling can inform policy design and crisis management schemes for urban scale smart infrastructure.