Urban income segregation, traditionally measured through static residential patterns, overlooks the dynamic social interactions shaping individual experiences. This study investigates the relationship between mobility patterns and experienced income segregation in 11 major US cities. Existing research predominantly focuses on residential segregation, neglecting the significant portion of daily life spent outside the home. This limitation hinders a comprehensive understanding of income segregation as experienced by individuals through daily encounters and interactions in diverse locations. The authors posit that individual mobility behavior, influenced by factors like exploration of new places and social interaction with different income groups, plays a crucial role in shaping experienced income segregation. This research aims to quantify this fine-grained income segregation and identify the behavioural and geographic factors associated with it. This is of significant importance because cities are increasingly becoming the center of social life, and understanding income segregation is crucial for building diverse, cohesive urban environments that foster sustainable development and positive socio-economic outcomes.

Existing research on urban income segregation often relies on static measures based on residential patterns, such as census data, and ignores the dynamic aspects of social interactions in various locations within a city. While several studies have attempted to approximate social interactions or exposure between income groups, this has been challenging to measure directly in the real world. Researchers often measure potential opportunities for interaction, focusing on physical proximity or exposure to different income groups in daily life. However, most city-dwellers spend considerable time outside their homes, and interactions occur in specific places, not just within large neighborhoods or census tracts. Previous studies have recognized that experienced income segregation can differ substantially from residential segregation, varying by location and time of day. This research emphasizes the need to move beyond traditional residential measures and incorporate individual mobility behaviors to gain a deeper understanding of income segregation as experienced by individuals in their daily routines.

This study utilizes high-resolution mobile phone location data from Cuebiq, encompassing 6 months of anonymized records for 4.5 million devices across 11 major U.S. cities. The data were supplemented with information from the Foursquare API for approximately 1.1 million verified venues. Each device was assigned a socio-economic status (SES) proxy based on the median household income of their inferred home area (Census Block Group). Individual visits to places lasting more than 5 minutes were extracted. The researchers introduce a metric, *S_a*, to quantify income segregation at the place level, ranging from 0 (full integration) to 1 (complete segregation). This metric measures the evenness of time spent by different income quartiles at a given venue. Similarly, a measure, *S_i*, was developed to quantify individual experienced income segregation based on the time spent in places with varying income segregation levels. To understand the relationship between place characteristics and income segregation, a regression model was employed. This model incorporated place characteristics (rating, price tier, category), catchment range (average travel distance to the place), and the median income of the block group. Geographic fixed effects were included to control for differences between areas within a city. To model individual mobility and income segregation, a Schelling extension of the Exploration and Preferential Return (EPR) model was used. This social-EPR model considers two parameters: place exploration (*σ_ε*), representing the tendency to visit new places, and social exploration (*σ_ρ*), reflecting the likelihood of visiting places where the individual's income group is a minority. Regression models were also used to examine the influence of various factors (lifestyles, geographical mobility, residential characteristics) on *σ_ε*, *σ_ρ*, and *S_i*.

The study reveals significant heterogeneity in place-level income segregation, with economically mixed and highly segregated places often in close proximity. Spatial correlation of place income segregation is low, even locally, unlike the higher correlation observed for block group median income. The distribution of place income segregation is surprisingly similar across different cities. Regression analysis showed that place category and catchment range are significant predictors of place income segregation, with places having higher average travel distances tending to be less segregated. However, these factors only account for a portion of the variance. Individual experienced income segregation is also heterogeneous and only weakly correlated with the income of the individual's home area. Experienced income segregation is not primarily driven by residential neighborhood but by the places individuals visit. A strong correlation exists between individual income segregation and the average income segregation of places visited. The social-EPR model accurately explains the patterns of individual visits and their experienced income segregation, highlighting the roles of place and social exploration. Place exploration (*σ_ε*) is primarily influenced by mobility behavioral variables (lifestyles, types of places visited), while social exploration (*σ_ρ*) is strongly associated with residential characteristics such as education level, employment, race, transportation mode, and poverty level. Both mobility behavioral factors and residential factors contribute significantly to overall individual income segregation, with the former accounting for approximately 55% of relative importance in the model. Only individuals exhibiting both high place and social exploration achieve economic integration.

This study provides novel insights into urban income segregation by demonstrating the significant role of individual mobility behaviors. The findings challenge the conventional focus on residential patterns alone and highlight the need for a more nuanced approach that incorporates dynamic interactions and location-specific segregation patterns. The social-EPR model offers a powerful tool for understanding how individual choices and behaviors, particularly place and social exploration, contribute to shaping experienced income segregation. This research underscores the importance of urban design and interventions that promote diverse and accessible spaces across the city, encouraging increased social and place exploration by individuals, potentially leading to more equitable interactions and reducing the experience of income segregation.

This research significantly advances our understanding of income segregation by integrating mobility data and behavior modeling. The study demonstrates that experienced income segregation is not merely a product of residential location but is intricately linked to individual mobility patterns and the characteristics of places visited. The social-EPR model successfully captures this relationship, emphasizing the roles of place and social exploration. Future research could explore the causal effects of urban interventions and design on social and place exploration, potentially revealing ways to promote greater economic integration in cities.

Several limitations exist. The study focuses on individuals with stable residences and excludes those with non-normative work patterns. The venues considered are limited to those accessible through the Foursquare API. The study relies on proxy measures of income and does not differentiate between different types of encounters. It focuses on income segregation and may not fully represent segregation based on other social dimensions. Finally, although suggestive, the study's findings are descriptive and do not imply causal relationships.