Abstract Summary
Since its inception in the late 1950s, Artificial Intelligence has been extensively researched, giving rise to many approaches within the field. Notably, the recent surge in demand for these new technologies has underscored the imperative of adapting them across diverse contexts, including architecture (Chaillou, 2021). However, as AI’s impact on architecture remains in its early stages, a comprehensive delineation of its benefits and drawbacks is yet to be achieved, particularly in a specialised field characterised by abundant data resources and emerging applications. Can Artificial Intelligence and Machine Learning, utilising the extensive urban architectural datasets, derive meaningful residential typologies from their morphological indicators? This research paper introduces a data-driven methodology designed to yield valuable insights for architects, urban designers, and planners. It emphasises the pivotal role of morphological analysis in comprehending the spatial characteristics and composition of residential buildings within urban landscapes. This approach leverages building footprint data and its inherent embedded information, focusing on a specific case study, the city of Amsterdam. The paper encompasses four distinct research domains: an exhaustive examination of accessible footprint data, advocating for high-quality and easily accessible data sources; the exploration of morphological geometry encoding, delving into shape definition methodologies; typology identification, where Machine Learning clustering techniques are investigated and deployed for optimal outcomes; and finally, the proposition of an analytical tool catering to remote analysts and individuals necessitating comprehensive urban and architectural information. Central to this work are the derived typologies, a systematic framework enabling the assessment of the heterogeneous building forms, sizes, and distributions that collectively shape our urban environments. These typologies yield promising outcomes, indicating that this approach could effortlessly bridge analyses across diverse domains of expertise and knowledge. This research illuminates the expanding potential of AI and Machine Learning in the field of architecture. The analytical framework presented here carries wide-reaching implications, envisioning a tool to develop our perception of the cities and buildings that make up our urban environments. Keywords: Data-driven; Morphology; Residential Buildings; Typologies; Analytical Tool.
