This paper proposes an innovative methodology for extracting and learning personal mobility patterns. The objective is to award daily commuters in a city with personalized and proactive recommendations, related with their mobility habits on a daily basis. In currently approaches, users have to explicitly provide their routes (origin, destination and date/time) to a routing engine in order to be notified about traffic events. The proposed approach goes beyond and learns daily mobility habits from the users, without the need to provide any information. The work presented here, is currently being addressed under the EU OPTIMUM project. Results achieved establish the basis for the formalization of the OPTIMUM domain knowledge on personal mobility patterns.