A Heuristic Approach to Generate Good-Quality Linked Data about Hydrography

Current Geographic Information is highly heterogeneous due to the diversity of producers and to their different needs. Geographical databases have different information structures, different levels of abstraction and scale, and are available in different natural languages. This is a major obstacle to overcome when generating good quality Linked Data from these databases. In this paper we describe how we generate Linked Data from heterogeneous hydrographical databases from various Spanish institutions. We provide a characterization of the types of heterogeneity found, based on existing semantic heterogeneity classifications, and describe a heuristic approach to deal with duplicity or co-reference problems.