Data, Information, Knowledge, Wisdom (DIKW)

The DIKW1 decomposition expresses a layering of context on top of raw observations.

Data points are attributes – components of facts. They are actuals (e.g., measured numbers) that can be connected to contexts.

Information is data tied to context, i.e. processed data (“history”) that reduces uncertainty wrt what, when, where, and who. Information connects attributes to compute entities – flocks of attributes that are in formation.

Knowledge is information tied to goals, i.e. actionable information (“know-how”) that reduces uncertainty wrt how. Knowledge connects entities to compute evaluations (e.g., which entities are most suitable for which goals?).

Wisdom is knowledge tied to action, i.e. applied knowledge that reduces uncertainty wrt why. Wisdom connects evaluations to compute decisions.

Data and information look to the past - information is data plus observed contexts. Knowledge and wisdom look to now and the future - knowledge is information plus goal contexts, and wisdom is knowledge plus action.

  1. Wikipedia overview. The ((what, when, where, who), how, why) decomposition was inspired by this Ontotext article↩︎