Understanding Urban Form - Hierarchical Morphotope Classification based on Gradually Loosening Spatial Restrictions

The physical structure of cities is under-studied due to data, methodological and theoretical limitations. We propose the hierarchical morphotope classification (HiMoC) - a scalable, hierarchical method for classifying urban form based on the concept of morphotopes — morphologically homogeneous areas. Using a bottom-up approach, it combines morphometric analysis with a bespoke clustering method called Gradually Loosening Spatial Restrictions (GLSR) to delineate and classify urban patterns into a taxonomic tree. We apply the method to 90 million buildings across six countries, to create a detailed, hierarchical urban typology while maintaining adaptability for various use cases. The taxonomy extends traditional classifications, offering deeper insights into urban structures and fostering reproducibility through open data and code