LODPeas: Like peas in a LOD (cloud)

We present LODPeas: a system for browsing entities that are found to share many things in common in an RDF dataset. The system first offers standard keyword search to locate a focus entity. Once a focus entity has been found, other entities that share a lot in common with it are displayed in a graph-based visualisation. The degree to which two entities have a lot in common—their level of concurrence—is scored by looking at attributes (property–value pairs) that they share: attributes that are shared by few other entities are given higher weight, and additional shared attributes imply a stronger score. LODPeas is designed to scale for billions of triples and is built in an (almost) entirely domain-agnostic fashion, built on top of the RDF standards themselves and not requiring any domain-specific input. Herein, we describe the functionality of LODPeas, how the system was built over the BTC’12, and discuss possible applications.