How to create a dataset#

A dataset holds a body of data together with the metadata that describes it: who made it, what it covers and how others may reuse it. Create it early, before the data exists, so that the work and the people behind it can be found.

Before you start#

  • You need to be signed in.

  • To place the dataset in a project, you need the Can edit level on that project. If you have none yet, create a project first.

Steps#

  1. In the header, select Create new, then Dataset. On a project’s page, the Add dataset button does the same. The Create a Dataset screen opens.

  2. In Name, type a descriptive name that says what the dataset holds. Others find the dataset by this name.

  3. In Project, choose the project the dataset belongs to. Leave it empty if the dataset belongs to none. You can attach it later on Edit details.

  4. In License, keep the portal’s default or choose another licence.

  5. In Visibility, choose Private or Public. Public is already selected. Choose Private if only your team should open the dataset’s page at first.

  6. Select Save & continue.

The result#

You land on the new dataset’s page, and a message says “Dataset successfully created.” The dataset lists you on its Contributors tab with the Can manage level and the roles Creator, Project Member and Contact Person.

Visibility covers the dataset’s page and metadata only. The dataset holds no data yet, and a dataset’s samples and measurements stay hidden from visitors until the dataset is published. See Dataset publication, DOIs and DataCite metadata.

To finish the dataset:

  1. Add an abstract and the methods on Descriptions.

  2. Add the collection period on Key dates.

  3. Credit the people who made it on Contributors.

The Ready to publish? card on the dataset page lists what is still missing. How to write metadata others can reuse says what to put in each field.