Populating the Semantic Web

The semantic web, or web of Linked Data, promises a transformation in the way information is combined, searched and presented. The dream is of an integrated planet-wide network of data, in the form of an RDF graph, that can be explored on our behalf by independent software agents. Despite the value of the prize, take-up has been slow and is only now gathering momentum. One barrier to entry is the complexity of converting information to the required form. This book examines that process in depth, using as an exemplar one of Scotland's national collections, the electronic data held by the Royal Commission on the Ancient and Historical Monuments of Scotland (RCAHMS). As well as being of intrinsic interest, this dataset provides an ideal test-bed for the key techniques: the conversion of structured relational databases to RDF, the automatic extraction of factual statements from free text documents, and the grounding of data against authoritative domain ontologies. The analysis of each process and the quantitative evaluation of SPARQL querying over the resulting 22 million triple graph should be of interest to anyone contemplating the publication of a large data collection as RDF.