The authors propose a topological analysis of large urban street networks based on acomputational and functional graph representation. This representation gives a functional view inwhich vertices represent named streets and edges represent street intersections. A range of graphmeasures, including street connectivity, average path length, and clustering coefficient, are computedfor structural analysis. In order to characterise different clustering degrees of streets in a streetnetwork they generalise the clustering coefficient to a k-clustering coefficient that takes into accountk neighbours. Based on validations applied to three cities, the authors show that large urban streetnetworks form small-world networks but exhibit no scale-free property.