Epigraphics: Message-Driven Infographics Authoring

TL;DR

Epigraphics introduces a text-first infographic authoring system that treats a key message (an epigraph) as a first-class driver for generating and syncing infographic assets, including static and animated visuals, graphics, color palettes, and data filters. The approach leverages large-language models and related tools to map message fragments to asset recommendations and supports rich between-asset interactions to maintain cohesive storytelling. Through gallery demonstrations, two case studies, and a 10-participant usability study, the paper shows that a message-sourced workflow standardizes content, fosters holistic design thinking, and accelerates rapid prototyping, while highlighting limitations in customization and layout automation. The work advances infographic authoring by enabling a controllable, generative, and modular pipeline that aligns asset generation with the designer’s intent, potentially enabling plugin integrations with existing design tools and broader adoption in data storytelling tasks.

Abstract

The message a designer wants to convey plays a pivotal role in directing the design of an infographic, yet most authoring workflows start with creating the visualizations or graphics first without gauging whether they fit the message. To address this gap, we propose Epigraphics, a web-based authoring system that treats an "epigraph" as the first-class object, and uses it to guide infographic asset creation, editing, and syncing. The system uses the text-based message to recommend visualizations, graphics, data filters, color palettes, and animations. It further supports between-asset interactions and fine-tuning such as recoloring, highlighting, and animation syncing that enhance the aesthetic cohesiveness of the assets. A gallery and case studies show that our system can produce infographics inspired by existing popular ones, and a task-based usability study with 10 designers show that a text-sourced workflow can standardize content, empower users to think more about the big picture, and facilitate rapid prototyping.

Figures (8)

• Figure 1: The complete pipeline from a text-based key message to infographic elements. It involves selecting text chunks from a key message (A), using these chunks to recommend different types of assets (B) such as visualizations, data filters, graphics, and color palettes, merging different combinations of the generated assets (C), and fine-tuning the configurations on a canvas (D).
• Figure 2: When the user brushes over a chunk of text, a pop-up with icons representing potential types of asset recommendations appears. After the user clicks on an icon, the asset is generated and automatically linked to the text chunk.
• Figure 3: A gallery of infographics created using Epigraphics with the corresponding epigraphs used to generate them. A mccrorie2016infographics, B fox2016every, C quealy2016sushi, and D lutz2014flight showcase recreations inspired by existing infographics, while E and F are originals based on open-source datasets. G zuniga2014what is also a recreation comparing what can be produced using our system (top) versus a traditional approach (bottom) with labels explaining their workflows for each asset type.
• Figure 4: The distribution of number of mouse clicks spent on asset generation and interacting with the canvas for each participant. All participants spelled the bulk of their clicks on manipulating assets on the canvas.
• Figure 5: A mapping of asset types that participants initiated recommendations for over time. Each dot is an instance where a participant requested a specific asset type using the system. Each light green rectangle spans half of the allotted time, 12.5 minutes, and is used to visually highlight when a majority of recommendations for that asset type occurred. Most participants focused on generating visualizations during the first half ($50\%$ of asset interactions), and graphics in the later half ($70\%$ of asset interactions).