On the Evolution of Semantic Warehouses: The Case of Global Record of Stocks and Fisheries

Marketakis, Yannis; Tzitzikas, Yannis; Gentile, Aureliano; van Niekerk, Bracken; Taconet, Marc

Semantic Warehouses integrate data from various sources for offering a unified view of the data and enabling the answering of queries which cannot be answered by the individual sources. However, such semantic warehouses have to be refreshed periodically as the underlying datasets change. This is a challenging requirement, not only because the mappings and transformations that were used for constructing the semantic warehouse can be invalidated, but also because additional information (not existing in the initial datasets) may have been added in the semantic warehouse, and such information needs to be preserved after every reconstruction. In this paper we focus on this particular problem in a real setting: the Global Record of Stocks and Fisheries, a semantic warehouse that integrates data about stocks and fisheries from various information systems. We propose and detail a process that can tackle these requirements and we report our experiences from implementing it.