Understanding neighborhood income segregation around the clock using mobile phone ambient population data

Abstract

Abstract This study examines the temporal changes in income segregation within the ambient population around the clock using mobile phone big data. It employs ordinal entropy, a metric suited for measuring segregation among ordered groups, to quantify the level of segregation among eight income groups within micro-geographic units throughout the 24-h period on a weekday and a weekend day in the urban core of Guangzhou, China. The study further decomposes daily segregation by location and time profile. We identify urban functions and neighborhood contexts relevant for income segregation and explore their temporal variation. Using group-based trajectory analysis, we classify daily segregation trends among 400 m urban grids into seven distinct trajectories for both weekday and weekend. Our findings confirm that segregation fluctuates constantly. The role of local urban functions, particularly retail, accommodation, and offices, and neighborhood context, such as the number of residents and the share of non-local migrants, exhibits a significant temporal rhythm. The seemingly convoluted 24-h segregation time series among urban grids follow just a few distinct trajectories with clear geographical patterns. There is limited variability at individual grids both over the course of a day and across days. Shifts across different trajectory types between weekday and weekend are rare. The dynamic daily segregation in the ambient population per se may be an enduring characteristic of neighborhoods and a real-time channel for neighborhood contextual influences, potentially fueling long-term residential segregation and neighborhood change.