Skip to article frontmatterSkip to article content
Site not loading correctly?

This may be due to an incorrect BASE_URL configuration. See the MyST Documentation for reference.

Session 1: JupyterLab + AI assistant as Data Steward Workbench

DSC ScaDS.AI, Leipzig University

Introducing data stewards to JupyterLab as a potential workbench for practical, reproducible workflows supporting a range of tasks (e.g. data intake and exploration, documentation, and DMP drafting), with an integrated AI assistant.

Starting from the basics, we set up a reproducible environment, learn our way around Jupyter notebooks, and then meet the bia-bob AI assistant - first for everyday coding help, then tailored into a data steward assistant that helps explore a real (raw) dataset and draft a plan, all while we read and verify what it produces.

You will learn to: work confidently in JupyterLab and Jupyter notebooks, understand why projects use isolated environments (uv / virtual environments), use the bia-bob assistant to generate and fix code, shape it with a custom system prompt, and apply it to explore data and draft documentation.

Notebooks

  1. 1_notebook-basics - JupyterLab and notebooks: cells, Markdown, running code, common pitfalls

  2. 2_ai-assistant-basics - the bia-bob assistant: prompting, generating and fixing code

  3. 3_ai-assistant-advanced - system prompts and a reusable Data Steward assistant

  4. 4_ds-workflow - a steward workflow: explore a dataset and draft a plan with the assistant

Leads into Session 2: Python Toolbelt for Data Stewards in JupyterLab.