A Relational Model of Neighborhood Mobility: The Role of Amenities and Cultural Alignment

TL;DR

This study addresses why some urban neighborhoods remain connected while others are isolated by proposing a relational model that emphasizes cultural alignment and amenity mix as soft infrastructure for mobility. It leverages two massive, cross-national datasets—Google Places-based co-visitation in the US and change-of-address mobility in Canada—together with measures of cultural scenes to predict inter-neighborhood connectivity via fixed-effects negative binomial models. The findings show that amenity-mix similarity and cultural-scene similarity robustly predict both short-term co-visitation and long-term residential mobility, with a reinforcing interaction between the two in both countries. The work highlights a cross-national, relational mechanism of urban cohesion and provides operational tools and datasets to inform urban policy and theory on segregation, mobility, and the spatial organization of cities.

Abstract

Why are some neighborhoods strongly connected while others remain isolated? Although standard explanations focus on demographics, economics, and geography, movement across the city may also depend on cultural styles and amenity mix. This study proposes a relational, cross-national model in which local culture and amenity mix alignment creates a "soft infrastructure" of urban mobility, i.e., symbolic cues and functional features that shape expectations about the character of places. Using ~650 million Google Places reviews to measure co-visitation between U.S. ZIP codes and ~30 million Canadian change-of-address to track residential mobility, results show that neighborhoods with similar cultural styles and amenities are significantly more connected. These effects persist even after controlling for race, income, education, politics, housing costs, and distance. Urban cohesion and segregation depend not only on who lives where or how far apart neighborhoods are, but on the shared cultural and material ecologies that structure movement across the city.

Figures (5)

• Figure 1: Comparative structural properties of urban mobility networks in the United States and Canada. The figure shows the networks studied and the distribution of four network measures across metropolitan areas in each country: (a) visualization of networks studied, the top right figure shows the Canada network (the zoomed part represents the Toronto area), and the top left one shows the U.S. network (the zoomed part represents the Chicago area), (b) number of edges, (c) average weighted degree, (d) average edge weight, and (e) weighted clustering coefficient. The U.S. network, constructed from over 650 million Google Places reviews linking establishments co-visited by the same users, exhibits greater density and higher connectivity, reflecting short-term, voluntary mobility captured through digital traces. In contrast, the Canadian network, derived from nearly 30 million officially registered address changes across three census years, is sparser, with weaker connections and lower clustering, representing long-term residential mobility.
• Figure 2: Urban connectivity networks in Chicago (top) and Toronto (bottom) reveal contrasting spatial mobility patterns. Each link represents a standardized connection (z-score of edge weight) between neighborhoods based on co-visitation (U.S.) or residential mobility (Canada). Chicago’s network revolves around the Loop and the North-South divide, while Toronto’s shows multiple hubs reflecting a more polycentric structure.
• Figure 3: Spearman correlation heatmaps show the relationships among demographic, voting, geographic, cultural, and amenity (mix) similarity measures, with connection strength as the focal variable. Results are shown for the United States (a) and Canada (b). Values in the range $[-0.05,+0.05)$ were omitted in the figure.
• Figure 4: Similarity in geography, amenities, culture, income, and race/visible-minority varies systematically across levels of inter-neighborhood connection strength. In Chicago (overall in 'a' and specific regions in 'b'), geographic similarity and amenity similarity dominate, while cultural and racial homophily increase mainly among the strongest ties. In Toronto (overall in 'c' and specific regions in 'd'), although all similarity measures rise with residential mobility, amenity and culture increase more among the strongest ties.
• Figure 5: Standardized coefficients and marginal effects from multivariate models showing the contributions of spatial, socioeconomic, voting, amenity, and cultural similarities (as well as their interactions) to inter-neighborhood connection strength. In the figure, we omit the geographic similarity coefficients because they are substantially larger than all other effects: U.S. $b = 0.975$ ($p<0.001$); Canada (undirected) $b = 0.917$ ($p<0.001$); Canada (directed) $b = 0.871$ ($p<0.001$). (a) Estimated standardized coefficients from the U.S. model, based on an undirected Google Places co-visitation network. (b) Log-link of predicted connection strength as a function of scene similarity across low, median, and high levels of amenity similarity in the U.S. (c) Standardized coefficients from Canadian residential-mobility models, comparing undirected and directed (origin–destination) specifications. (d) Log-link of predicted connection strength in Canada as scene similarity varies across amenity-similarity percentiles, shown for both undirected and directed models. Across all countries and model specifications, amenity-mix alignment and cultural similarity emerge as substantial and consistent predictors of inter-neighbourhood connection strength, underscoring their role as core drivers of urban cohesion beyond demographic or spatial proximity.