Identification of Land Use Mix Using Point-Based Geospatial Data in Urban Areas

Abstract

Identifying land use mix (LUM) in urban areas is challenging, often requiring extensive human intervention and fieldwork. Accurate classification of LUM is crucial for various disciplines, including urban planning, urban economics, and public health. This study addresses this need by employing Voronoi triangulation and an entropy-based LUM formula using point-based geospatial data collected from publicly available sources. The methodology was tested in two distinct urban settings: Ankara and Kadıköy. Ankara, the capital city, provides a large and diverse urban environment, while Kadıköy, a district in Istanbul known for its dynamic urban life, offers a contrasting scenario. Results were analyzed concerning local spatial autocorrelation and point of interest (POI) intensity. The comparative analysis demonstrated that the approach performs well across different urban contexts, with improved results observed in Kadıköy due to its higher density of mixed-use development. Specifically, we managed to identify mixed land use areas with an accuracy of up to 78% and an F1-score of 83% in urban regions. These findings highlight the robustness and applicability of our approach in diverse urban environments, providing valuable insights for city planners and policymakers in optimizing the allocation of urban resources and enhancing land use efficiency.

Keywords

• land use mix
• point of interest data
• Voronoi triangulation
• local spatial autocorrelation
• geospatial data mining