The purpose of an integration-oriented ontology is to provide a conceptualization of a domain of interest for automating the data integration of an evolving and heterogeneous set of sources using Semantic Web technologies. It links domain concepts to each of the underlying data sources via schema mappings. Data analysts, who are domain experts but not necessarily have technical data management skills, pose ontology-mediated queries over the conceptualization, which are automatically translated to the appropriate query language for the sources at hand. Following well-established rules when designing schema mappings allows to automate the process of query rewriting and execution.