The streets’ network shapes the urban layout that structures the city space and delineates homogeneous urban areas. Structural analysis of the streets’ network is very useful because it usually produces additional semantic information (about urban dynamics). This allows the enrichment of spatial data. The topological and/or geometric characteristics of the network can be analyzed using some graph theory algorithms. The aim of this work is to perform some characterization of the urban fabrics using exclusively the streets’ network as input data. Indeed, the structural properties of the streets’ network create several effects such as continuity, separation, centrality, proximity, connectivity, symmetry, and adjacency into the urban space. We work on an urban primal graph where streets are considered as edges and streets’ intersections are considered as nodes. On the basis of this real urban graph, we implement two additional artificial graphs. The first one is a “Delaunay triangulated network” and the second one is a “minimum spanning tree”. These graphs correspond to two extreme network morphologies delimiting a morphological continuum. The three graphs are then analyzed using shortest path algorithms. They all have been processed on both a topological and a geometric (weighting each edge by its length) ways so as to extract some structural and multi-centrality indicators (closeness and betweenness centralities). The results can produce semantic information on the structure of the urban fabrics. They allow us to emphasize some specific morphological urban structures (organic areas, suburban tree structures) and some urban salient features (main roads and ring roads).