Data Standards
The data standards of the Velocity Interoperability Network govern how the departments and network services of the Network publish data through the open data catalogue. They exist so that data from different publishers can be read with the same tools, joined on the same identifiers and understood from the same metadata. The Data Office will not list a dataset that does not meet them.
The first edition of the standards was adopted on 6 September 2026. The Office reviews them annually, in consultation with the publishing bodies and with users of the catalogue, and publishes each revision in the Newsroom with a period of at least three months before it takes effect. Datasets published under a previous edition are brought into conformity at their next scheduled update.
The standards. The six standards below apply to every dataset in the catalogue.
- Formats
- Every dataset is published as comma-separated values encoded in UTF-8 with a single header row. Datasets with a nested or hierarchical structure are also published as machine-readable objects, and datasets describing places are also published in a geographic object format. Proprietary formats are not accepted.
- Metadata
- Every dataset is accompanied by a title, a description of its contents and coverage, the name of the publishing body, the licence, the update frequency, the date of first publication, the date of the most recent update, and a description of every column giving its name, type, unit and permitted values.
- Identifiers
- Member organisations are identified by the reference assigned in the register of member organisations; departments and network services by the reference assigned in the directory of departments; and sites and places by the reference assigned in the register of public facilities. Publishing bodies must use these references rather than names so that records can be joined across datasets.
- Dates and times
- Dates are recorded as year, month and day in that order, separated by hyphens. Times are recorded on the twenty-four hour clock with the offset from universal time. Periods are recorded with an explicit start and end rather than a label.
- Missing and suppressed values
- A value that is not known is left empty. A value that has been suppressed to prevent the identification of an individual is marked with a suppression code, and the metadata states the rule under which suppression was applied.
- Versions and corrections
- A dataset that is updated on a schedule replaces the previous release, and the previous release remains available for twelve months. A correction to a published dataset is recorded against the dataset with the date and a description of what changed.
Approved formats. The table below sets out the formats in which data may be published and the circumstances in which each is required.
| Format | Required for | Requirements |
|---|---|---|
| Comma-separated values | Every dataset | UTF-8, single header row, one record per line, fields quoted where they contain commas or line breaks |
| Machine-readable objects | Datasets with nested or hierarchical structure | One document per dataset, or one document per record for series updated continuously |
| Geographic objects | Datasets describing sites, places or areas | Coordinates recorded to the Network's common reference frame, stated in the metadata |
| Plain text | Documentation and column descriptions | UTF-8; supplied alongside the dataset, never within it |
Reference datasets. Three datasets in the catalogue serve as the reference datasets of the Network: the register of member organisations, maintained by the Member Registry; the directory of departments and network services, maintained by the Ministry; and the register of public facilities, maintained by the Interior Department. The identifiers they assign are permanent: an identifier is never reused once assigned, and a body or site that ceases to exist keeps its identifier with an end date recorded against it.
Publishing a dataset. Departments and network services wishing to publish a dataset through the catalogue follow the steps below. The Office's Registry and Standards Unit advises publishing bodies at every stage.
Step 1: Confirm that the data can be released
The publishing body checks the classification of the holding in the register of data holdings. Only holdings classified as open, or as open in aggregated form once aggregated, may be published in the catalogue.
Step 2: Prepare the dataset to the standards
The data is exported in the required formats, using the Network's identifiers, with dates and times in the required form and any suppression applied and documented.
Step 3: Write the metadata
The publishing body completes the metadata for the dataset, including a description of every column. The Office supplies a template and will advise on any point of difficulty.
Step 4: Submit the dataset to the Office
The dataset and metadata are sent to the Office by the officer responsible for the holding, with a proposed publication date and update frequency.
Step 5: Checks by the Office
The Office checks the dataset against the standards and the metadata against the data, normally within five working days, and returns any points for correction to the publishing body.
Step 6: Publication
The Office lists the dataset in the catalogue on the agreed date, adds it to the register of data holdings as published and announces significant releases in the Newsroom. Subsequent updates are supplied by the publishing body on the agreed schedule.
Publishing bodies may write to contact@data.gov.vin for the metadata template, for advice on applying the standards to a particular holding or to propose an amendment to the standards. For the terms under which published data may be reused, see Licensing and Reuse; for the classification of holdings, see the Register of Data Holdings.