Technology
Our primary assumption in CommonPlace is that the various data providers can expose their datasets as RDF graphs. When this is not the case, the project can support them through, with resources and knowhow, to transform their datasets.
The CommonPlace knowledge graph is a Linked Open Data (LOD) collection that is an aggregation of many datasets. Although in an open world assumption of LOD all factual observations together form a single, massive graph or semantic web, in practice users will only be interested in certain elements of the graph. We enable this by introducing the concept of a Lens: a selection that can come from one, but frequently originates from many datasets to provide a unique and singular view on the graph for a user. Lenses allow the user to select one interpretation among many different historical observations. For example, the previously mentioned data object ‘Winston Churchill’ may be described as follows: A <Parish Record> records that <Winston Churchill> was <baptised> on <April 18th, 1620> at <St. Dunstan’s-in-the-West in Fleet Street, London>. However not all data providers are able to offer RDF triples with this degree of precision. WikiData, for example, asserts that <Winston Churchill> was <born> (note: not <baptised>) on <April 18th, 1620>. The CommonPlace data infrastructure will not attempt to pick out one assertion amongst many as the ‘best’ interpretation. Instead, the user can either select from preconfigured lenses of a certain precision or specific source; or configure a personal lens by applying an appropriate threshold for data quality, inference, and trust to certain data providers.