Adding Relationships to the iiindex

As of last week, it is now possible to add relationships between different datasets in the iiindex.

Why add relationships?

Part of the interest in starting the index came from a wish to better understand the citation landscape of datasets used by members of the I3, at the very least to establish which versions of popular datasets served as the basis for widely-used resources in the community.

An awareness of how particular datasets are derived, combined with a . However, obtaining this information is tricky and time-consuming, as most researchers don’t use a formal structure for data citation.

As such, we wanted to find a good way to crowdsource this information into the index. In a similar way to other kinds of metadata — like notes and code samples — this is a field that benefits from being primarily edited by humans, not bots, and as such should be simple and easy to edit.

How to add a relationship

At present, the simplest way to add a relationship is using the site itself. On each dataset page (found either by using the search on the homepage, or by browsing the full list), it’s possible to reveal a form by clicking the ‘+’ symbol next to “add relationship“ in the pages’ infobox. By selecting the dataset you’d like to relate it to, using the search and dropdown menu, and then selecting the type of relationship it has to that resource, click ‘submit’ and the relationship (and its inverse) will be added to the database.

From submission, it will take 1-2 minutes for the site to update with the new relationships.

It’s also possible to add relationships by editing the markdown files in the Github repository directly, though this does not automatically generate the inverse relationship.

Why does it take a minute to update the site?

Like all changes to the index, it is managed on the back end using Github and Google Sheets, meaning that every change is versioned, recorded, and editable by anyone. Because of this, whenever an edit is made the site is re-compiled, and