This site collects the materials for the workshop we’re giving at the DGPs 2026 conference in Luxembourg. If you’re a participant, you shouldn’t need much beyond this page. If you’re not — welcome anyway. Everything here, including the notebooks, is meant to be followed along at your own pace.
By the end of the workshop, you’ll be able to (1) tell apart the three main types of LLM-based text classifiers and know when each one fits your use case, (2) run all three on real, open-ended survey text using Hugging Face and Google Colab, and (3) evaluate whether a classifier’s output is good enough to trust for your research question.
All materials, including the notebooks, are also available from the GitHub repository if you’d like your own copy or want to work through the exercises again later.