Context data-driven approach refers to the process of collecting and storing context data from a wide range of context sources as sensors and Web services. This approach differs from existing context-aware applications, where context models and applications are closely related and ignore how the context is derived from sources and interpreted. In this paper, we propose a context data model based on semantic Web technologies for ubiquitous computing applications. This model facilitates the specification of context data to be easy interpreted by applications and services. In addition, this model is supported by existing communications protocols. Therefore, our context data model is applied to promote mining of context data towards service adaptation in ubiquitous computing systems.