Find
Belgian public data is spread across federal, regional, local and specialist portals. That fragmentation makes sense organisationally, but it makes discovery harder for everyone else.
We are working on bringing more of those sources together in one searchable catalogue and making them easier to browse.
We are also exposing the catalogue through MCP so AI assistants can help people discover relevant datasets without first knowing which portal to search.
And discovery should not depend on somebody actively visiting a data catalogue. Through Belgian Open Data on social media, we are also trying to surface interesting datasets, examples and findings to a wider audience.
These are different ways of contributing to the same goal: making public data easier to find.
Trust
Open data should not be treated as trustworthy simply because it has been published.
People need context.
Who published it? When was it updated? What licence applies? Is there a contact? Does the download still work? Is the file actually what the catalogue says it is?
We are working on making more of that evidence visible and automating checks where that is useful and realistic.
Those checks do not prove that a dataset is correct. They give people more information to make their own judgement and make problems easier to notice and report.
Understand
To understand data, you need to see it.
That means being able to look beyond a title and description and inspect what is actually inside: the columns, values, structure, gaps and relationships.
We are experimenting with previews, profiling and browser-based querying with DuckDB, and we want to explore lower-friction ways to compare and combine datasets.
Quizzes and explainers are another part of that effort. They can help people become more comfortable reading public data critically and recognising its limitations.