The use of Intelligent Transport Systems has created a new possibility for obtaining large volumes of data that can modify its operation. In the case of public transport systems, the information from the Automatic Fare Collection Systems plays an important role, since it can obtain information on the use of smart cards, data on the place of entry, value paid and the user's own characteristics. However, there is information derived from the trip that cannot be obtained directly from the data from the use of smart cards, such as the reason for the trip, is that from the validation information of the SITP trunk system ( TransMilenio), data processing and mining is carried out, travel matrices are built from chaining and the application of heuristics for their inference, in order to develop multinomial logit-type models, which obtain the reason for the trip, which are compared with the origin of destination household survey in Bogotá, D.C.