Data Integration for Querying Geospatial Sources

Geospatial data management is fundamental for many applications including land use planning and transportation and is critical in emergency management. However, geospatial data are distributed, complex, and heterogeneous due to being independently developed by various levels of government and the private sector. Until now, the formulation of expressive queries on geospatial data, which contrast with simple keyword-based queries, requires both user expertise and a great deal of manual intervention to determine the mappings between concepts in potentially dozens of data sources. In this paper, we describe an ontology-based approach to the problem of data integration, specifically focusing on the issue of query processing in a heterogeneous setting. Our contributions include a mechanism for metadata representation, an ontology alignment process, and a sound query rewriting algorithm for answering queries across distributed geospatial data sources. We demonstrate the practical impact of our approach in land use applications, which are exemplary of the extreme heterogeneity of data. We are leveraging current and emerging Semantic Web standards and tools for modeling, storing, and processing data. Our contribution to geospatial data integration is significant because new data sources can be added with relatively little effort, thus allowing for data manipulation and querying to extend seamlessly to the new data sources.