Sensor data is becoming far more available thanks to the growth in both sensor systems and Internet of Things devices. Much of the value of sensor data comes from examining trends that occur over long timescales, ranging from hours to years. However, making use of data a long time after it has been collected has significant implications for the data-handling systems used to manage it. In particular, the data must be contextualised into the environment in which it was collected to avoid misleading (and potentially dangerous) mis-interpretation. We apply data warehousing techniques to develop an extensible model to capture contextual metadata alongside sensor datasets, and show how this can be used to support the analysis of datasets long after collection. We present our baseline reference framework for sensor context and derive multidimensional schemata representing different modelling and analysis scenarios. Finally, we exercise the model with two case studies.