A data mining based methodology for the multidimensional study of public open spaces

Abstract

Public open spaces can only be apprehended from multiple simultaneous perspectives. Urban morphology traditional descriptive methods have recognized limitations in relating the polymorphic and polysemantic nature of these spaces’ attributes, derived from the different standpoints on their formal, historical and geographic idiosyncrasies. Identities and similarities may be disclosed by multivariate statistical analysis and data mining techniques by studying the relations between formal and intangible spatial properties in a multidimensional space. In an ongoing PhD research project we outline a method for the synchronic analysis and classification of the public open spaces, departing from a corpus of 126 Portuguese urban squares, whose analysis is intended to interactively (re)define it. Part of the work done so far is presented: (i) firming the concepts, criteria and attributes to extract; (ii) adaptation and/or creation of new analytical methods and tools; and (iii) research on multivariate analysis, data mining and data visualization techniques.