Key points are not available for this paper at this time.
This study investigates whether spa settlements can be analytically interpreted through a relational spascape framework that reveals structural and configurational patterns beyond conventional typological classifications. In the context of increasing interest in therapeutic landscapes and heritage-sensitive development, spa settlements represent complex spatial systems shaped by the interplay of natural resources, urban form, and socio-cultural practices, yet they remain insufficiently understood through existing analytical models. The methodology is based on a structured analytical design combining three urbanization dimensions (material transformation, territorial regulation, and everyday life) with six thematic fields, operationalized through graded cross-affiliation scoring. The empirical research is conducted on a sample of 12 spa settlements in Serbia, selected to reflect diverse geographical, morphological, and developmental conditions. Statistical calibration was performed using Principal Component Analysis (PCA) and hierarchical clustering to identify underlying structural relationships and configurational groupings. The results indicate that spa settlements operate as multi-affiliated relational entities rather than fixed typologies, exhibiting dimension-specific structural logics and forming distinct configurational families depending on the analytical perspective applied. PCA reveals differentiated internal structures across dimensions, while clustering confirms the absence of a single stable typology. The findings support a relational understanding of spa settlements as dynamic spatial systems characterized by shifting alignments of material, regulatory, and experiential factors. Beyond the Serbian context, the study offers a transferable methodological framework that connects qualitative urban interpretation with quantitative spatial analysis, contributing to heritage-sensitive planning, territorial governance, and the management of spa systems as relational clusters.
Milovanović et al. (Tue,) studied this question.