Power Centrality as a relational measure of urban hierarchy : testing the splintering urbanism theory with social media data from Santiago de Chile

Abstract

Power Centrality, a measure of node importance within a network, is borrowed from the field of Social Network Analysis and applied to the assessment of Urban Hierarchy. Based on the overlaps of human activity between places, Power Centrality is tried as a method for measuring a particular feedback property: How well connected are places to other well connected places. In this research Power Centrality is used to assess a recent model of Urban Structure: The Splintering Urbanism Theory of Graham and Marvin (2001). This theory posits that the contemporary city is a fragmented agglomeration of isolated urban pieces, where distant but valuable fragments are highly connected between them, bypassing their less valuable surroundings. The causal explanation provided by Graham and Marvin is centered on their concept of premium networks: Networks customized for valuable (users in terms of income or power). The reach of this theory is assessed by studying the case of a mass transit system in a developing country: The Metro or subway of Santiago de Chile. The spatial hypothesis of Graham and Marvin is tested empirically through the use of the Power Centrality Measure, applied to a dataset of 242.000 twitter statuses generated by Metro users, while the causal explanation is evaluated by comparing the results with an unbiased sample of 110,000 statuses. Power Centrality allowed the identification of central locations that by standard measures of spatial concentration would have remained undetected. Furthermore, the results evidenced how Metro could be acting as a mass public bypass that connects these emergent centralities, challenging the concept of premium networks posited by Graham and Marvin.