Providing personalized offers, and services in general, for the users of a system requires perceiving the context in which the users' preferences are rooted. Accordingly, context modeling is becoming a relevant issue and an expanding research field. Moreover, the frequent changes of context may induce a change in the current preferences, thus appropriate learning methods should be employed for the system to adapt automatically. In this work, we introduce a methodology based on the so-called Context Dimension Tree---a model for representing the possible contexts in the very first stages of Application Design---as well as an appropriate conceptual architecture, to build a recommender system for travelers.