This paper develops a classification of mobile interactions based on contextual information relevant to the mobile device user in journeys. Context-aware systems can be used to reduce the stress involved, support users in their activities and increase utility of travel time. But context is often portrayed as real, stable and structured, which can limit the value of applications as they lack dynamics and relevancy. This paper aims to classify mobile interactions in journeys by adopting an alternative view of context. It is argued that sensing less contextual information can be more valuable providing the most relevant information to the user can be identified. Context is explored using qualitative approaches that investigate user interactions during end-to-end journeys. The resulting classification serves as a basis for understanding mobile interactions and it assists designers and HCI practitioners to develop improved context-aware application.