Metabase: Wikibase for the Wikimedia Movement

An interview with Wikimedia Sverige

In this interview we hear from the Metabase project, developed by Wikimedia Sverige as part of the Content Partnerships Hub initiative. Metabase is a platform designed to collect and organize information about the Wikimedia movement’s activities and resources in a linked, structured format. Launched officially in 2024, after two years of intermittent development, Metabase was built on Wikibase, leveraging its features for ease of data input, querying and integration with Wikidata. The goal of the project is to create a collaborative space where affiliates and volunteers can access and share valuable data, fostering knowledge exchange and capacity building across the Wikimedia community. 

To start, please provide an overview of the Metabase project, its history and its objectives.

Metabase is a wiki – a Wikibase instance – where information about the activities done and resources produced by the Wikimedia movement are stored in a linked, structured form. It is being developed by Wikimedia Sverige as part of the Content Partnerships Hub initiative. Our intention is to build a platform where all this information can be collected and accessed by everyone, affiliates and volunteers alike.

The first item was created in October 2022, and we have been working on the project intermittently since then. In the summer of 2024, we felt the project was ready to be officially introduced to the community: we published a white paper (which elaborates on all the issues in this text) and spoke about Metabase at Wikimania in Katowice.

Being faced with an empty Wikibase and the complete freedom to decide what it should be filled with is quite overwhelming. After all, we were not just making a toy for ourselves, but a platform that would hopefully be used by other members of the Wikimedia movement.

Wikimedia Sverige

What’s the role and significance of Wikibase features in organizing your data?

Our most important requirement for the platform was that it should support linked, structured data. Ease of both data input and querying was also important. Since we’re Wikimedians, Wikibase was an obvious choice – it looks and feels like Wikidata. The development of the Wikibase.cloud platform lowered the threshold even more, since we didn’t have to concern ourselves with installing and maintaining our own Wikibase instance. It’s like having our own tiny Wikidata!

How do you think Wikibase will enhance findability and searchability of your data?

The data model, which we’re familiar with after years of using and contributing to Wikidata, is simple yet flexible. It allows us to break down our data into small, meaningful pieces. And SPARQL, obviously, makes the magic happen, allowing users to find whatever they need. As it allows us to run federated queries, we can collate our data with that in Wikidata, which in turn means we do not have to duplicate data from Wikidata if it’s not absolutely necessary, allowing our data model to remain pretty shallow. For example, a Wikimedia affiliate in Metabase is just an organization, a city is just a location – but since they also have Wikidata QID’s, we can run a query to find conferences arranged by organizations that are part of the Wikimedia movement and plot their locations on a map.

What challenges did you encounter while working with Wikibase in the Metabase project, and how did you address them?

We’ve had some technical challenges related to Wikibase.cloud – even though it looks like Wikidata, there are some major differences. For one, it lacks support for gadgets and user scripts. You don’t realize how much you depend on things like Merge, LabelLister, DuplicateReferences, MoveClaim and Constraints until you have to do without them! It would also be nice to be able to modify our editing experience with custom CSS and JavaScript.

The biggest challenge, however, was of an ontological nature. It turns out that being faced with an empty Wikibase and the complete freedom to decide what it should be filled with is quite overwhelming. After all, we were not just making a toy for ourselves, but a platform that would hopefully be used by other members of the Wikimedia movement. New contributors needed to be able to understand both the scope of the project and how to model the data.

In order to address this issue, we really took our time at first. The Metabase team at WMSE is small – two persons in the beginning, with another member joining us later on, all experienced Wikidatans. We had a content-driven modelling workflow, focusing on the questions we wanted Metabase to help us answer and working our way back from there to figure out how the data should be structured. Our properties and modeling practices are very close to Wikidata’s for a reason – it should be as easy as possible for Wikimedians to understand the ontology and start contributing.

What practices do you employ to ensure the sustainability of your Wikibase infrastructure?

This is a very big question. In fact, we were asked it when speaking about Metabase at Wikimania 2024 – if people are to invest time into learning and contributing to our platform, they need to feel their work is meaningful and will benefit others in the future. We feel safe hosting the data in Wikibase.cloud, and we appreciate the fact that it can be exported in a standardized format.

There’s also the question of, shall we say, organizational sustainability. The bigger Metabase becomes, the more important it will be to curate the content, clean up any vandalism and take care of the community (not that we expect our contributors to misbehave!). Right now Metabase has three admins, the WMSE team members. But it might not be enough very soon. When we have active users with a track record of great contributions, we hope they will be willing to help out with admin tasks and helping newcomers. We would love to see a community around our project!

In your view, what role could Wikibase play in the wider Wikimedia Movement? How might other Wikimedia chapters and user groups benefit from Wikibase and the Linked Open Data web?

As mentioned in the beginning, Metabase is an attempt to address a real problem in our movement. We do not know whether it will be successful – it’s way too early to tell. Even though we have been working on the project on and off for two years, we only started getting contributions from non-WMSE people very recently. The way we see it, Wikimania 2024 could be seen as a sort of “real” launch for the project, since that’s when we finally felt we were ready to invite the movement to contribute – not only with data, but also with ideas and discussions (and criticism!).

Our hope for Metabase is that it will become an important part of our shared knowledge ecosystem, a place for affiliates and volunteers alike to learn from each other. Metabase should make it easier for everyone to share and discover capacity-building resources, which in turn will help everyone build up their skills faster and do more great Wikimedia things. 

---

In this interview we heard from:

• Alicia Fagerving, developer, Wikimedia Sverige
• André Costa, Deputy Executive Director, Wikimedia Sverige