A Facet is an aspect or type of entry within a knowledge graph entity. Examples of facets include the location of an organization’s headquarters, or the number of individuals performing a certain role at the organization. Faceting is similar to filtering with some differences. Filters allow a user to specify some criteria that may or may not match any entities to be included or excluded. Faceting provides a view of the prevalence or non-prevalence of certain attributes within entities who are returned.
Diffbot’s Knowledge Graph™ allows for faceted search on up to 1,000 KG entities at one time. Examples of a filter at work in Diffbot’s Knowledge Graph would include a search for all software companies located in San Francisco with less than 50 employees. This would return a list of organizational entities that meet this criteria. On the other hand, a faceted search could include the above search with the facet parameter of industry. This would return a view of what industries the search result companies are in.
In the case of interlinked entities such as Knowledge Graphs, faceting does not exclude the ability to dive more deeply into individual entity entries. Rather, faceting can provide a high level view of the characteristics of entities to inform future analysis.