Mapping signs of urban exclusion: machine learning-based clustering of spatial accessibility in tourism cities

Yüksel T. G., Kızılırmak İ.

INTERNATIONAL JOURNAL OF TOURISM CITIES, cilt.1, ss.1-15, 2026 (Scopus)

Özet

Tourism cities often contain complex physical barriers that may spatially constrain the accessibility of tourists with disabilities, highlighting the need for machine learning (ML)-based approaches. Grounded in the Social Model of Disability, this study investigates the spatial clustering of accessibility across Istanbul’s Historic Peninsula, a UNESCO World Heritage site, by bridging social perspectives with algorithmic spatial analysis and assessing disparities between attractions and amenities within the 6A destination framework. Using unsupervised ML, the study analyses 3,186 tourism-related Points of Interest (POIs). Four digitally available wheelchair accessibility features were aggregated into an accessibility score. Following the CRISP-DM, non-parametric variance analysis was utilised to assess categorical disparities, while the DBSCAN algorithm was applied to map spatial density patterns. Findings reveal that potentially exclusionary clusters concentrate in specific urban areas, highlighting a disconnect between attractions and amenities. Specifically, 96% of digitally invisible and visible-inaccessible accommodation facilities formed dense, interconnected clusters. By algorithmically mapping these clusters, the study reveals the signs of exclusion in urban areas for tourists with disabilities based on digitally available wheelchair accessibility data, rather than direct observations of accessibility barriers or tourist experiences. Findings provide urban planners with a data-driven approach for equitable smart governance.