Towards Collective Storytelling: Investigating Audience Annotations in Data Visualizations

TL;DR

This paper tackles how audience annotations can transform data visualizations from data-centric displays into environments for collective storytelling. By designing coronaMoments, the authors compare three reading conditions that vary annotation origin (audience vs. author) and integration (embedded vs. separated) across two studies. They find that embedded audience annotations most effectively engage readers, provide personal and social context, and guide interpretation through social traces, while also fostering empathy and reflection. The work highlights methodological considerations, design implications for participatory visualization, and the potential to democratize discourse around data, alongside cautions about moderation, topic fatigue, and trust. Overall, it advances understanding of how crowd-sourced, context-rich narratives can augment data-driven storytelling and public discourse.

Abstract

This work investigates personal perspectives in visualization annotations as devices for collective data-driven storytelling. Inspired by existing efforts in critical cartography, we show how people share personal memories in a visualization of COVID-19 data and how comments by other visualization readers influence the reading and understanding of visualizations. Analyzing interaction logs, reader surveys, visualization annotations, and interviews, we find that reader annotations help other viewers relate to other people's stories and reflect on their own experiences. Further, we found that annotations embedded directly into the visualization can serve as social traces guiding through a visualization and help readers contextualize their own stories. With that, they supersede the attention paid to data encodings and become the main focal point of the visualization.

Figures (4)

• Figure 1: Our design space spans two dimensions: annotation origin and annotation integration. Each combination is illustrated by examples described in the Background, for each example the annotations are highlighted orange (audience annotations) or blue (author annotations): (A) annotations from the audience as investigated in social analysis heer2007voyagers or participatory mapping kirby2021queering; (B) embedded annotations curated by visualization authors as common in data-driven storytelling\ref{['storytelling']}; (C) separate sections for comments from the audience in visualization-related articles or social media posts hullman2015contentkauer2021public, and (D) author's comments on the data that are separate from the visualization as often done in data journalism
• Figure 2: Three views of the coronaMoments platform: (left) Embedded annotations collected from the audience, (middle) embedded annotations added by the author, (right) separated annotations collected from the audience.
• Figure 3: Across all conditions we compare, how many unique visitors viewed it (1), how long each session lasted (2), how many annotations they created (3), how often per session they enlarged annotation bubbles to read them (4), how many times per session they clicked the "show random moment" button (5), how often per session they filtered the view by selecting a hashtag (6) or a country (7).
• Figure 4: Survey responses of people who read annotations (red) and after they submitted own annotations (blue).