Diurnal Patterns and Social Ecology: Throughput in Urban Spaces

• In Event: 4350 - Computational Sociology (Co-sponsored with Mathematical Sociology)

Abstract

While we understand that city life is strongly patterned, specifically in 24 hour increments, the forces behind this patterning are not as well understood. I examine one of the forces that has been suggested as a predictor and driver of daily life, social ecology. Measuring this concept as combinations of institutions, I use quantitative methods including OLS regression to predict the patterns of daily life. Drawing on literature from social ecology, geography and sociology, I use this regression model to understand how these institutions could drive the 24 hour pattern of life. This project uses cell phone data from the city of Milan across a period of two months to examine this question, leveraging methods drawn from signal processing to refine diurnal signals. I find that while social ecology is a strong predictor of the amount of social activity in a city, other latent forces are certainly pertinent in the production of this pattern of life. To better understand why the OLS model can only predict part of the pattern, I use random forests to better explain why the temporal phase of this activity is difficult to estimate. These results have implications for organizers of social movements, criminologists, and epidemiology, as it allows people to better understand the movement of people around a city.

Authors

• Loring Thomas, University of California, Irvine

• Carter T. Butts, University of California-Irvine