Identifying urban development clusters in southern Poland using the DBSCAN algorithm for spatial analysis

What the paper says

Urban sprawl is a critical issue for people’s lives, our economies, our cities and our natural environments. This article explores urban development patterns as they are manifested in spatio-structural urban models in metropolitan areas (cluster, linear or dispersed) vis-à-vis evidence of sprawl. We focus on four metropolitan regions in southern Poland: the Wrocław metropolitan area, the Górnośląsko-Zagłębiowska metropolis, the Kraków metropolitan area and the Rzeszów metropolitan area. We use the GIS heat map tool to identify spatial models in our study areas and we isolate clusters of continuous compact development (and the spatial relationship of these clusters to urban boundaries) using the DBSCAN (density-based spatial clustering of applications with noise) algorithm. Findings suggest that most buildings (60 per cent to 65 per cent) tend to form clusters outside city boundaries across the four metropolitan areas studied; however, an exception is the Górnośląsko-Zagłębiowska metropolis, where clustering predominantly occurs within city limits. In addition, 20 per cent to 35 per cent of buildings in metropolitan areas across our study areas are not part of clusters, but the Górnośląsko-Zagłębiowska metropolis is again an exception, with only 5 per cent of buildings outside clusters. The research leads to four recommendations: promote cluster models for sustainable development, taking into account historical and policy influences; limit sprawl by adopting cluster model strategies in suitable urban zones; and pursue monocentric agglomerations, with the goal of sustainable urbanisation; cluster analysis with the DBSCAN algorithm determines more realistic urban development boundaries than the heat map tool. We propose action-oriented implementation strategies to reach these aims.