Writing SPARQL queries with GitHub Copilot

Format: 1.5 hour online workshop

Date and Time: 12/12/2025, 2-3.30pm UK time (UTC+00:00)

Convenor: Duncan Hay, University College London

As many people working with Linked Open Data will attest, writing SPARQL queries is hard. Happily, Large Language Models, especially tools such as GitHub Copilot, are very good it. In this workshop, you will be introduced to how to set up and use Visual Studio Code and Github Copilot to write SPARQL queries and test them against the Tools of Knowledge database of scientific instrument makers.

After the demo and a chance to run your own queries, we will conduct a facilitated discussion around possible use cases, and the advantages and disadvantages of this approach. Questions addressed will include:

• How might this be used in different work and research contexts?
• How else might these tools be used to work with Linked Open Data? (for example GraphRAG and related methods)
• How good are the queries that Copilot generates?
• What gets lost when you offload your thinking to Copilot in this way?

Prerequisites

Before the session, you should have:

• Visual Studio Code installed on your computer https://code.visualstudio.com/
• A GitHub Account. If you are affiliated with a University you should be able to get an Education account which gives you very generous usage limits with Github Copilot. See here: https://github.com/education/teachers or here: https://github.com/education/students.
• Git installed on your computer https://desktop.github.com/download/
• The GitHub Copilot extension installed in Visual Studio Code https://marketplace.visualstudio.com/items?itemName=GitHub.copilot