Multidimensional analysis of public open spaces. Urban morphology, parametric modelling and data mining

João Ventura LOPES *1

Alexandra PAIO *1

José Nuno BEIRÃO *2

Eliana M. PINHO *3

Luís NUNES *1

*1 Instituo Universitário de Lisboa

*2 Faculdade de Arquitetura da Universidade de Lisboa

*3 Faculdade de Arquitetura da Universidade do Porto

Abstract

Public open spaces can only be grasped from a simultaneous view of their attributes. Given its many shapes, functions and uses, we propose to address the limitations of traditional-descriptive urban morphology in dealing with this multidimensionality. Through the implementation of quantitative methods, based on multivariate statistical analysis, the relations between formal and intangible spatial attributes, the identity and an unambiguous classification of public open spaces is specified. The combination of several theories, methods and tools of spatial and formal analysis with data mining allows us to tackle the elusive nature of the subject of study, creating new bottom-up knowledge from the very structure of the data itself.

The proposed generic multidimensional method of synchronous analysis and classification of public open spaces is implemented in a corpus of Portuguese squares. Through computational processing of their digital models and in situ observations, we retrieve a heterogeneous set of formal, spatial, functional, environmental, social and perceptual-cognitive attributes, building an open-ended and scalable database for subsequent multivariate analysis. This approach enables: (1) to correlate quantitative and qualitative/performative attributes, supporting a stronger evidence-based design; (2) to propose an alternative classification of the corpus, indicating possible new lines of research, including contemporary spaces that defy formal classifications; (3) to comparatively analyse the descriptive-morphotypological and the configurational-structural approaches to urban analysis and design (typo vs population thinking); and (4) to propose a synthesis capable of combining the most significant contributions of each to the understanding and improvement of the quality of public open spaces.

(This abstract refers to an ongoing PhD research project supported by the FCT SFRH/BD/95148/2013 PhD grant.)