Statistics and Data Science Using Agentic AI

Statistics and Data Science Using Agentic AI

The course helps researchers across the natural sciences and engineering do faster and more rigorous data analysis with agentic AI tools, such as Claude and OpenAI's Codex. It is practical throughout: we take the kinds of data problems researchers actually face — plotting, hypothesis testing, regression modelling and other machine learning models — and show how to do them more efficiently. It starts with the basics of effective prompting, moves on to Skills and instruction files for repeatable results, and finishes with an agentic workflow you build yourself in Python or R. Along the way we explain how agentic AI actually works — it is simpler than many people realise — which is what tells you where its limits and capabilities lie. The emphasis throughout is on correctness: context engineering, where the agent is supplied with a context object so that outputs are grounded in the correct data and background and are therefore reproducible and explainable, together with test agents that check your work and the manual checks that good data practice asks for. The final two weeks cover reproducibility, sharing your work and open science, and the ethical aspects and rules around AI use in publications.

By the end, each participant has built their own subject-specific statistics and data science agent to use in their day-to-day work. For PhD students, a report on their work will be used to grade the course.

Prerequisites

Familiarity with at least one quantitative research method. Some experience with Python or R is helpful but not required. The course is open to anyone at Uppsala University.

When, where and how?

8 weeks, 5 credits (HP for PhD Students, check with supervisor/FUAP).

Online videos uploaded each week, plus a two-hour drop-in workshop at Ångström each week (also available on Zoom). Help will also be provided over Slack.

Taught by David Sumpter (Department of Information Technology) with one PhD teaching assistant. Teaching is in English.

Course start 5th October 2026. Register here

For questions contact David Sumpter