A dataset has a set of named entities. Other forms such as different spelling and synonyms are mapped to one of the entities as aliases or marked as invalid.

Statistics: 43 Entities, 7534 Aliases (745 marked invalid).


Aliases are added via the API, by matching names against this endpoint:

http://nomenklatura.okfnlabs.org/uk25k-column-names/lookup?name=foo

An OpenRefine reconciliation endpoint is available as well (read-only):

http://nomenklatura.okfnlabs.org/uk25k-column-names/reconcile?query=foo