Mapping Mismatch: Perceived vs. Statistical Diversity in Urban Activity Spaces

The study of segregation in activity spaces is experiencing renewed attention, fueled by the increasing availability of mobility data from mobile phones and social media. These new data sources allow for high-resolution analysis of how people move through cities and interact across diverse neighborhoods. A key goal behind this work is to inform the design of inclusive spaces that foster intergroup contact and social cohesion. However, much of this research has focused on statistical diversity (typically measured through census data, mobility traces, or indices that capture the distribution and co-presence of social groups) while giving less consideration to how diversity is perceived by those who use these spaces.This study bridges that gap by combining GIS-based mobility modeling with on-street survey data. Using aggregated mobility data from ~90,000 mobile phone users in Auckland, New Zealand, we estimated ethnic mobility flows and computed spatial diversity indices for residents and visitors across six neighborhoods. In parallel, we conducted 249 on-street surveys in those neighborhoods to capture perceptions of diversity.Our findings reveal only a weak correlation between statistical and perceived diversity, highlighting that individuals do not simply mirror demographic compositions in their perceptions. Instead, perceived diversity is more strongly predicted by self-reported intergroup contact (both within the specific neighborhood and across the city) and by affective factors such as space affinity and attitudes toward outgroup members. These results underscore the importance of integrating both statistical and perceptual data to understand diversity in urban environments. Designing for inclusion, we argue, must account for both who is present in a space and how that space is experienced.