We consider graphs as models of different sorts of transportation networks, both at national level: highways and plane routes, and at city level: subway and bus systems. Graph theory techniques and matrix analysis consider centrality functions which define a ranking, depending only on the topological structure, among the vertices of the graph. These functions provide strategic information for the resolution of various problems. Closeness centrality is used in the study of the efficiency of the transportation networks. Spectral centrality (the local values of an eigenvector associated to the spectral radius) is used in the study of the coherence of the transportation networks. We introduce the concept of coherence as a measure of how a transportation network structure meets the needs of users. We use real data of the mexican transportation systems and draw conclusions on their performance.