'Eyes' on the Street: What Public Camera Feed Data Can Teach Us About Civic and Political Behavior

Abstract:

I introduce a new type of data that is increasingly available to social science researchers: geo-located real-time pedestrian traffic counts.  Data on foot traffic, singular in its precision and temporal coverage, comes from a private company that uses a network of live video camera feeds and computer vision technology to calculate traffic levels in real time.  After discussing the data structure and limitations, I present several social science applications that describe the relationship between space and ‘civic-ness’ in a highly walkable urban setting.  First,  I show how foot traffic counts can be used to study large events and gatherings by examining pedestrian flows during a recent political protest.   Second, I use spatial kriging to interpolate hourly foot traffic in Manhattan, which is then overlaid with geo-located New York City 311 reports, a proxy for citizen engagement with local government. I explore how certain non-emergency issues are more likely to be reported under varying conditions. Next, I bring data on incidents of major crime and show that the most highly trafficked blocks are also the most dangerous. Finally, I conclude with ideas for the future use of the data. 

Pdf

Last updated on 08/26/2016