A semantic approach to data translation: A case study of environmental observations data

To facilitate the exchange of environmental observations, efforts have been made to develop standardised markup languages for describing and transmitting data from multiple sources. Along with this is often a need to translate data from different formats or vocabularies to these languages. In this paper, we focus on the problem of translating data encoded in spreadsheets to an XML-based standardised exchange language. We describe the issues with data that have to be resolved. We present a solution that relies on an ontology capturing semantic gaps between data and the target language. We show how to develop such an ontology and use it to mediate translation through a real scenario where water resources data have to be translated to a standard data transfer format. In particular, we provide declarative mapping formalisms for representing relationships between spreadsheets, ontologies, and XML schemas, and give algorithms for processing mappings. We have implemented our approach in AdHoc, an ontology-mediated spreadsheet-to-XML translation tool, and showed its effectiveness with real environmental observations data.