Metadata is data that describes other data. It provides context that makes data understandable, findable, and usable.
Think of a library book. The book itself is the data. The library catalog entry — title, author, subject, publication date, call number — is the metadata. Without the catalog, finding the right book in a library of thousands would be nearly impossible. The same principle applies to datasets in any organization.
Types of Metadata
Descriptive Metadata
Describes what a dataset contains and helps people find it. Includes: title, description, keywords, creator, creation date, geographic coverage, temporal coverage, and subject categories. This is the metadata you see on an open data portal when you search for a dataset.
Structural Metadata
Describes how a dataset is organized. Includes: column names, data types, relationships between tables, file format, encoding, and schema definition. A data dictionary is a form of structural metadata. Without it, a CSV file full of cryptic column codes is nearly unusable.
Administrative Metadata
Supports the management of a dataset. Includes: owner, steward, access rights, licence, retention schedule, version history, and update frequency. Administrative metadata answers governance questions: who owns this data, who can use it, and how long should it be kept?
Technical Metadata
Describes the technical characteristics of a dataset. Includes: file size, number of records, coordinate reference system (for spatial data), and software requirements. Technical metadata is essential for data engineers who need to process or integrate datasets.
Many metadata standards exist for different domains. Dublin Core is a simple, widely used standard for descriptive metadata. ISO 19115 is the standard for geographic metadata. DCAT (Data Catalog Vocabulary) is used for open data portals. Using standards makes metadata interoperable across systems — a dataset described with DCAT can be automatically harvested by any DCAT-compatible portal.
Metadata in Practice
A municipal GIS team publishes a dataset of park locations. Good metadata tells users: what coordinate system is used, when the data was last updated, which department owns it, what licence applies, and what each field means. Without this context, users cannot confidently use the data — they do not know if it is current, authoritative, or compatible with their tools.
Key Takeaways
- Metadata is data about data — it provides context and meaning
- Four types: descriptive, structural, administrative, and technical
- Standards like Dublin Core and DCAT enable interoperability
- Good metadata makes data findable, understandable, and trustworthy
Learn the concrete value of metadata in Why Metadata Matters.