Open & FAIR data

FAIR means Findable, Accessible, Interoperable, and Reusable. Open data and FAIR data overlap, but they are not the same: data can be private under a valid access policy and still follow FAIR practices, while a public file with no metadata may be difficult to reuse.

openJII provides useful building blocks; it does not make a dataset FAIR automatically or certify FAIR compliance.

A practical checklist

Findable

• Give the experiment and datasets descriptive names.
• Add structured metadata for subjects, treatments, locations, conditions, and units.
• Preserve experiment, protocol, workbook, device, and export identifiers in your research record.

Accessible

• Choose public visibility or document the embargo and access conditions.
• Set collaborator roles deliberately: admin or member.
• Keep a human-readable contact and citation route with the released dataset.

Interoperable

• Export in a documented standard format such as CSV, JSON, Parquet, or Excel.
• Use stable column names, explicit units, standard vocabularies, and ISO dates.
• Explain nested values, raw traces, flags, missing values, and any contributor-anonymization choice.

Reusable

• Record the protocol, workbook version, hardware, calibration context, and transformation or macro used.
• Include quality annotations and known limitations.
• State a license and citation expectation outside the platform if the project requires one; the current experiment create flow does not assign a data license for you.

Package a release

Export the intended table, save the metadata and methods alongside it, and keep the export date and row-count checks. Where possible, publish analysis code and an environment lockfile or container recipe so another researcher can reproduce the transformation.

For field definitions, see the Glossary. For technical provenance through the processing pipeline, see Architecture & Pipeline.