Can generative AI create good data visualisations? – Nicola Rennie: “This blog post compares the performance of ChatGPT, Claude, Copilot, and Gemini when presented with a generic request to visualise a dataset. Nicola Rennie October 8, 2025. “Generative AI tools have been around for a little while now, and they’re mostly advertised as being able to help you with writing, coding, and reasoning. Different AI tools are designed for different types of tasks, so I wanted to see how some of the most commonly used tools compare to each other when used for data visualisation. I’m going to compare four different generative AI tools – ChatGPT, Claude, Copilot, and Gemini. Using the free tier of each tool, I’m going to give them two different datasets with a prompt to create a chart. Prompts – I didn’t want to use well-known datasets like penguins or titanic since there are so many examples of charts of those datasets already in existence. I wanted to see how well they would do with new(ish) data. If you’ve been following these blog posts for a little while, it’s probably not a surprise that I turned to TidyTuesday to find some data!

Prompt 1: weather forecast data – The first dataset used gives information on weather forecast accuracy from the USA National Weather Service. It contains data on the observed and predicted (at 12, 24, 36, and 48 hours before) temperature and precipitation levels for different cities in the USA, between January 2021 and June 2022…”