Andmestike

Andmestike is the Estonian plural of andmestik, a term used to describe datasets—collections of data records organized for analysis, reporting, or operational use. Datasets can vary in size, structure, and subject matter, and may be prepared for specific tasks such as statistical analysis, machine learning, or archival preservation.

A dataset typically consists of data values arranged in records and fields, or records and attributes. They can be structured (for example, in tables with defined columns), semi-structured (such as JSON or XML), or unstructured (like photographs, audio files, or free text). Each dataset often includes metadata that describes its content, provenance, methods of collection, quality indicators, licensing, and update frequency. Data dictionaries, schemas, and lineage information help users interpret and trust the data.

Formats and standards are diverse and often chosen for interoperability. Common formats include CSV, JSON, XML, Parquet, and relational database schemas. Metadata standards such as Dublin Core, ISO 19115, or DCAT-AP support discovery and interoperability, while domain-specific schemas provide structure for particular fields or industries.

Management practices for andmestike emphasize governance, access, and quality. This includes defining ownership, usage licenses (for example, open data licenses), privacy protections and anonymization where needed, version control, and long-term preservation. The lifecycle of a dataset involves creation, curation, validation, sharing, and potential deprecation.

Open and well-documented andmestike enable reproducibility, collaboration, and informed decision-making across research, government, business, and civil society.