datat

Datat is a term used in information science and data management to refer to a unit of data that is packaged with its context. Unlike the general term data or a dataset, a datat is typically described as a self-contained data artifact that includes the data payload plus metadata, provenance, and governance information. The concept has emerged in discussions of data catalogs, data governance, and data marketplaces, where it helps distinguish a portable, contract-like data unit from a plain collection of records.

Core characteristics of a datat include a documented data payload, accompanying metadata (such as schema, data quality, and lineage), a provenance trail that records its origin and transformations, a persistent identifier for stable reference, and access controls and licensing terms. Datats are intended to be format-agnostic, supporting structured, semi-structured, and unstructured data, and they are typically versioned to support reproducibility.

Datats relate to but are not identical with data stacks, datasets, or data products. A datat may be the primitive unit in a data catalog, whereas a dataset is often a larger collection or slice used for analysis. In practice, the term is sometimes used informally and with varying definitions, which can lead to ambiguity between organizations.

Typical applications include data catalogs, data governance frameworks, data marketplaces, reproducible research workflows, and machine learning pipelines that require traceable data provenance and licensing.

Because the term is not standardized, its precise meaning can differ by context. Users should refer to explicit metadata schemas and governance policies to ensure interoperability when datats are exchanged across systems.