Our paper aims to apply a non-conventional Social Network Analysis to a network generated from an initial sorting of data with mere order. The aim is to test whether and how one relationship set 'conveniently' generated by the ranking is able to provide a coherent system of interactions with possible practical and theoretical utility. We consider a system of city distributions where we focus on ¿Ranking-Based Ties¿ social networks, considering a system of 'objects' of the same class that would be able to interact. The objects in question are ordered according to their 'performance' in a particular ranking. According to a survey developed by Global Language Monitor, there is a ranked number for every capital that could be considered important among the world's most fashionable cities. Using the rank of scoring average we construct a fashion cities network in order to analyze the dynamics and the topology of the system following a gradual process of clustering or densification of relationships consistent with key network statistics analysis.