Even important connections are not always meaningful: on the use of a polarisation measure in a typology of european cities in air transport networks

摘要

This paper aims to refine existing approaches for classifying nodes in spatial networks. Apart from generally used variables measuring (i) the total connectivity of a node and (ii) the spatial make-up of this total connectivity, we argue that a third indicator should be taken into account when describing the structure of a spatial network, namely, (iii) the degree of 'polarisation' in a node's connectivity configuration. We measure 'polarisation' by applying multiple linkage analysis, a technique that allows determining the number of meaningful connections for each node in the network. In this paper, we illustrate this approach by clustering 65 European cities based on their insertion in European air transport networks in 2005. The number of meaningful flows of a city is hereby considered alongside both traditional measures of connectivity. The results of the cluster analysis suggest that including a measure of polarisation may indeed lead to a more satisfying typology.