The results presented in this paper might help to support this problem. A method is proposed that enumerates all possible decompositions of a given path in a graph, in polynomial time. This allows to easily calculate for every vertex the occurrence frequency as a splitVertex in a minimum path decomposition. This is a measure for the probability that the vertex was used as an intermediate destination by travelers. It contains information about the network (topology, path travel cost) and about the revealed routes. This information can be derived from big data (GPS traces). It quantifies the appropriateness of a vertex to act as a landmark useful for route construction. The level of abstraction is much lower than the one associated with the MRI in Kazagli and Bierlaire (2014) but the information can be derived automatically.