Wednesday 27 January 2027 | 11:30 - 13:00 I Workshop Room

A hands-on session on using AI agents to speed up data visualization work,  without sacrificing rigor.

We start with the non-negotiable principles (visual encoding, chart choice, clarity, honesty) that any good graph needs, then show how to offload the repetitive parts (styling, iteration, code generation) to AI agents so you spend your time on decisions, not syntax. 

Data and AI

Learning outcomes

By the end of the workshop, participants will be able to: 

  • Create a mental checklist for "is this a good graph" independent of tools 

  • Build a repeatable workflow: principles → prompt → critique → refine

  • Make practical prompting patterns for getting AI to produce publication-quality plots faster 

Who should attend

This workshop is designed for: 

  • Researchers/analysts who already make plots (R/Python, ggplot/matplotlib) but want to cut iteration time,  since the AI part assumes you are able to judge whether the output is good. 

     

Workshop outline

  • Why "AI speed" without foundations = fast garbage. Core principles: right chart for the data, visual hierarchy, color/typography discipline, avoiding chartjunk 

  • Live examples: good vs. bad plots, what separates them 

  • How AI agents fit in: prompting for viz code, iterating on plots conversationally, generating/critiquing against a style guide 

  • Live demo: build a plot from raw data to polished figure using an AI agent, applying the earlier principles as prompts/constraints 

  • Hands-on: participants try a prompt themselves

Workshop Leader

Loading
Speaker profile image for Loukas Theodosiou

Loukas Theodosiou

Senior Bioinformatics Scientist
Max Planck Institute for Evolutionary Biology

**Please register your interest in being considered for a workshop place when completing your festival ticket registration.**