Autopoietic Features of the Urban Body’s Elements – Similarity studies on network elements’ attributes

eCAADe 2016: Complexity & Simplicity

2016, University of Oulu, Finland – Conference Proceedings

Adopting a methodology towards an understanding of the urban body whose elements include material ingredients (nodes) and their inscribed properties (attributes) through the construction of a network configuration, the autopoietic dynamics of the urban body emerge, as properties of things -urban areas inscribed on the Attic peninsula in Greece, revealed through the application of autopoietic -network functions, measuring qualities which reflect the network structure.

We call autopoietic those nodes’ features which emerge from the inner structure of the urban body and are quantified through the application of community detection (Blondel et al., 2008, Rosvall et al., 2009), centrality (betweenness centrality, degree centrality, closeness centrality, Newman, 2010) and other algorithms which capture the network qualitative features of the network nodes and reveal the real-time self-adaptation of the parts and components of the urban body, on the basis of the successive introduction of nodes with different autopoietic features (Figure 1).

We focus on the autopoietic properties of the network’s nodes as they are quantified through the application of certain autopoietic processes which reveal the nodes’ interrelations, translating the relationships between things into dimensions (attributes’ values) of elements. Through the quantified expressions of the inner dynamics of the  network configurations the autopoietic properties emerge. These define the ‘species’ of the elements according to their inner relations (network attributes in general), the parts of the urban body as articulated fractions/sub-networks (centrality measures), as well as the components of the urban body as areas of stronger interrelated elements (community detection), where the areas’ boundary changes according to the desired threshold of the interrelated strength. The emergence of those features of the urban body elements which are inherent in the network structure and the generation of the urban body parts and components through the application of autopoietic processes, reveals the inner structure of the urban body, as a reflection of its inherent dynamics and made possible through the examination of the urban body as a network configuration. Following this methodology and in order to reveal further relationships between network and urban body structures, supra-local nodes are regarded as multidimensional objects and the attributes emerging from the network configuration appear as a set of values (dimensions), allowing for a clustering methodology of objects and their attributes, based on similarity studies and made possible through the application of  different distance and linkage functions (Hastie, Tibshirani, Friedman, 2003). The application of  the distance and linkage functions results in a data tree where nodes, as well as their autopoietic attributes, are visualized in a color mosaic which captures the proximity of data values to each attribute’s lowest and highest extreme values and in an hierarchical tree, where the position of each element reflects its relational similarity following a gradual procedure of  selection of two sets of elements based on the strength and the direction of their relation. This approach suggests the transition from a clear perception of an object/urban site to the ambiguity of multiple properties of things, a set of data values rendered in the form of scattergrams, color mosaics and dendograms. Among the possibilities originating from this approach is the understanding of the concurrent alterations taking place in a given network construction among all node’s characteristics with the shift of one property, a kind of structural coupling between body and environment (Maturana,  2002).