Semantic querying in Earth observation data cubes

Earth observation (EO) data cubes have revolutionized the way large volumes of EO data can be stored, accessed, and processed. However, users coming from application domains outside of traditional EO research still face some significant technical barriers when querying an EO data cube with the aim to infer knew knowledge about real world entities and events. They have to interpret EO data in order to give them meaning, which is an ill-posed problem that requires advanced expertise in the field of EO analytics. We propose a semantic querying framework in which users query the EO data cube through an ontology, rather than accessing the data values themselves. The ontology formalizes symbolic representations of real-world entities and events, which are mapped to data values in the EO data cube through a mapping component formulated by an EO expert. This takes away the need for users to be aware of the EO data and how to interpret them, and therefore lowers the technical barriers to extract valuable information from EO data. We implemented a proof-of-concept of our approach as an open-source Python package.