Using AI & Python for Data Visualization

A hands-on 2-hour workshop introducing seaborn — Python's most approachable statistical graphics library — alongside an AI coding assistant workflow. Participants learn just enough seaborn to direct and verify an AI (Gemini in Colab, ChatGPT, Claude, or any tool they already use) as it drafts charts. All examples use the clean datasets that ship inside seaborn (penguins, tips, flights, titanic), so there are no files to upload.

In this 2-hour in-person session, you'll learn how to:

• Use seaborn for the three families of plots — relational (scatterplot, lineplot), distribution (histplot, boxplot), and categorical (barplot, countplot)
• Layer multiple variables into one figure with semantic mappings (hue, size, style, col) — seaborn's grammar-of-graphics approach to mapping data columns to visual properties
• Apply five core visualization principles (chart choice, reducing clutter, intentional color, labeling, avoiding misleading charts)
• Work with an AI coding agent (Gemini in Colab) using the describe → generate → run → read-the-error → iterate loop
• Write effective visualization prompts (dataset & columns → goal → constraints → chart type)
• Recognize and catch the common failure modes of AI-generated charts (hallucinated functions, wrong chart types, misleading scales, altered data)
• Apply the full workflow to your own data or research question in the open-work section

What to know about this session:

This training is hands-on; come ready to code alongside the instructor. No prior plotting experience is required — if you're comfortable with basic Python and familiar with pandas DataFrames, you're ready. We use Google Colab (which has Gemini built in — no separate AI account or API key needed). Details about technical setup will be emailed to registrants a few days before the course begins. Please note that registration is required for this event.