Billion Triple Challenge 2010: Namespace
High Performance Semantic Factoring of Giga-Scale Semantic Graph Databases
To understand the relationships between the sources which generated the BTC dataset, we explore a summary metric for data linking among namespaces.
We call triples "linked" when two or more of the subject, predicate, or object map to different fully-qualified domain names.
Using our Cray XMT, we extracted domain names from each entity and computed a census of both the domains and linked triples. With this data, we crafted an interactive tool which visually presents the relationships between these namespaces.
Summary census information: