DataTales: Investigating the Use of Large Language Models for Authoring Data-Driven Articles



Nicole Sultanum, Arjun Srinivasan

 Room: 104

2023-10-26T22:27:00ZGMT-0600Change your timezone on the schedule page
2023-10-26T22:27:00Z

Exemplar figure, described by caption below — User Interface of DataTales on the background, and DataTales logo on the foreground. User interface features a stacked bar chart titled "America's favorite & least favorite months of the year", the contents of a data story generated for this chart, and a list of other generated stories. A mouse cursor hovers over a text passage, and the corresponding data elements mentioned in the passage are highlighted in the chart.

Fast forward

Full Video

Keywords

large language models; data-driven articles; data-driven authoring; data stories

Abstract

Authoring data-driven articles is a complex process requiring authors to not only analyze data for insights but also craft a cohesive narrative that effectively communicates the insights. Text generation capabilities of contemporary large language models (LLMs) present an opportunity to assist the authoring of data-driven articles and expedite the writing process. In this work, we investigate the feasibility and perceived value of leveraging LLMs to support authors of data-driven articles. We designed a prototype system, DataTales, that leverages a LLM to generate textual narratives accompanying a given chart. Using DataTales as a design probe, we conducted a qualitative study with 11 professionals to evaluate the concept, from which we distilled affordances and opportunities to further integrate LLMs as valuable data-driven article authoring assistants.