Leveraging Text-Chart Links to Support Authoring of Data-Driven Articles with VizFlow, CHI 2021
Nicole Sultanum, Fanny Chevalier, Zoya Bylinskii, Zhicheng Liu
Data-driven articles — i.e., articles featuring text and supporting charts — play a key role in communicating information to the public. New storytelling formats like scrollytelling apply compelling dynamics to these articles to help walk readers through complex insights, but are challenging to craft. In this work, we investigate ways to support authors of data-driven articles using such storytelling forms via a text-chart linking strategy. From formative interviews with 6 authors and an assessment of 43 scrollytelling stories, we built VizFlow, a prototype system that uses text-chart links to support a range of dynamic layouts. We validate our text-chart linking approach via an authoring study with 12 participants using VizFlow, and a reading study with 24 participants comparing versions of the same article with different VizFlow intervention levels. Assessments showed our approach enabled a rapid and expressive authoring experience, and informed key design recommendations for future efforts in the space.