Datasets
Overview
How datasets work in Envision Portal.
A dataset is the primary object in Envision Portal. It groups together files, metadata, documentation, and access controls into a single versioned unit.
Dataset States
Every dataset exists in one of two states:
| State | Description |
|---|---|
draft | Work in progress. Only team members can view and edit. Files are stored in draft storage. |
published | Publicly visible (or discoverable). Files are moved to published storage. A DOI is assigned. |
Versioning
Each dataset has a canonical ID that remains constant across all versions. When you publish a new version, the canonical ID stays the same and the version number increments. This allows stable external references via the canonical ID while still tracking history.
Dataset Sections
A dataset is made up of several sections that you fill in progressively:
- General information - Title, description, version, type
- Metadata - Contributors, identifiers, related identifiers, access rights, data management, team
- Study metadata - Clinical trial details (optional, for clinical datasets)
- Healthsheet - Data quality documentation
- Files - Uploaded data files
- README - Auto-generated from metadata, editable before publishing
- Changelog - Version history, generated and editable before publishing
Navigation
When you open a dataset in the app at /app/datasets/[datasetId], the sidebar provides access to all sections. Publishing is a separate workflow accessible from the dataset overview page.