Using Large Language Models to Generate Engaging Captions for Data Visualizations



Ashley Liew, Klaus Mueller

 View presentation:2022-10-16T20:03:00ZGMT-0600Change your timezone on the schedule page
2022-10-16T20:03:00Z

Exemplar figure, described by caption below — Creating compelling captions for data visualizations has been a longstanding challenge. Visualization researchers are typically untrained in journalistic reporting and hence the captions placed below data visualizations tend to be not overly engaging and rather just stick to basic data observations. We explore the opportunities offered by the newly emerging crop of large language models (LLM) which use sophisticated deep learning to produce human-like prose. We ask, can these powerful software devices be purposed to produce engaging captions for generic data visualizations like a scatterplot. We report on first experiments using the popular LLM GPT-3 and deliver some promising results.

The live footage of the talk, including the Q&A, can be viewed on the session page, NLVIZ: Opening, Keynote and Paper Session I.

Abstract