Semantically Enriching VGI in Support of Implicit Feedback Analysis

In recent years, the proliferation of Volunteered Geographic Information (VGI) has enabled many Internet users to contribute to the construction of rich and increasingly complex spatial datasets. This growth of geo-referenced information and the often loose semantic structure of such data have resulted in spatial information overload. For this reason, a semantic gap has emerged between unstructured geo-spatial datasets and high-level ontological concepts. Filling this semantic gap can help reduce spatial information overload, therefore facilitating both user interactions and the analysis of such interaction. Implicit Feedback analysis is the focus of our work. In this paper we address this problem by proposing a system that executes spatial discovery queries. Our system combines a semantically-rich and spatially-poor ontology (DBpedia) with a spatially-rich and semantically-poor VGI dataset (OpenStreetMap). This technique differs from existing ones, such as the aggregated dataset LinkedGeoData, as it is focused on user interest analysis and takes map scale into account. System architecture, functionality and preliminary results gathered about the system performance are discussed.