A multiscale clustering of the urban morphology for use in quantitative models

Abstract

Geo-spatial data available to researchers and practitioners has increased substantially over the last decades, offering new opportunities for analyzing and characterizing locations and their spatial environment in an objective way. Various studies in the field of urban planning and design have given recommendations on “good urban form”, suggesting that especially characteristics of the urban morphology inform the quality of locations and additionally influence spatial behavior. While we find a growing number of quantitative spatial models such as hedonic price models, location choice models, or applied machine learning algorithms, characteristics describing the urban morphology are still rarely reported in these. One reason can be found in the limited knowledge on how to characterize the urban morphology best. In this chapter, we address this deficit and define attributes that are suitable to characterize the urban morphology in quantitative means at different scales. We show how these can be processed from a data model that is simple enough to allow for reproducibility in most study areas and process them for the case study of Switzerland. Finally, we use the attributes to define urban typologies through clustering methods. These are compared on their outcome, their consistency, and interpretation.