datumot

Datumot is a theoretical unit in information studies that measures the informational value of a single data observation within a defined analysis context. Unlike conventional data size metrics measured in bytes, a datumot is intended to reflect how much a given data point reduces uncertainty about a target variable or decision. As such, datumots are context-dependent and not standardized across studies.

Origin and usage ideas for the term have emerged in contemporary discussions about data value, provenance, and decision-making under uncertainty. It is primarily used in thought experiments and exploratory analyses to distinguish raw data volume from actionable information, rather than as a universal measurement.

Formal interpretation centers on information gain. In probabilistic terms, a datumot can be thought of as the expected information gain or reduction in entropy produced by incorporating that observation, averaged over a specified prior and model class. Calculating datumots requires defining the prior distribution, the target quantity, and the set of models under consideration, making the value contingent on these choices.

Applications and relevance appear in debates about data pricing, data sharing incentives, and data quality assessment, where the same data size can yield different practical value depending on context. The concept emphasizes value over volume, but its lack of standardization limits direct comparability across analyses or communities.

See also: information content, entropy, mutual information, data provenance, data value, data ethics.