This study examines methods for modelling relationships between urban form patterns and socio-demographic characteristics using spatially explicit architecture. By comparing linear and nonlinear dimensionality reduction techniques alongside geographically weighted prediction models, we analyse the interplay between population and morphological patterns in Czechia. The study combines a morphometric classification of over 4 million buildings with detailed 2021 Census data at the basic settlement unit level. This methodological framework evaluates the effectiveness of various modelling approaches contributing to our understanding of the relationship between people and their built environment.