Background

Effective data visualization is important for any field that uses data. Data visualizations allow audiences to perceive data patterns much more efficiently and accurately than written or verbal descriptions. When designed appropriately, data visualizations can enhance audience members’ trust in the data and give them confidence in decisions based on that data.

Best practices in data visualization have largely been developed by designers, statisticians and computer scientists. Surprisingly, there is still little evidence about whether humans perceive, interpret and interact with data visualizations in the way that best practices assume. This gap in how we are hypothesized to process data visualizations and how we actually process them limits our ability to predict what data visualizations will be most effective or most misleading. This can deeply impact society when data visualizations are critical for applications like making medical decisions or determining whether an artificial intelligence model is fair. 

Project Description

This project team will leverage cognitive science tools to evaluate some of the most popular data visualization “best practices” and adjudicate between ways of applying them to individual data visualizations. The team will use eye-tracking analysis software, cognitive surveys and qualitative interviews to test whether popular visualization practices have the intended impact on their audience’s perception, evaluation and understanding of data visualizations in different contexts.

Team members will then develop visualizations that intentionally challenge the efficacy of best visualization practices and engage 2-3 data visualization experts to make predictions about how audience members will respond to those visualizations. The team will collect eye-tracking, facial expression, questionnaire and qualitative interview data to test those predictions. Team members will share results from both phases in a public-facing blog to bring visibility to the team’s work and inspire others in the data visualization field to incorporate empirical testing into their practice and thinking.

Generated data will be used as preliminary evidence for whether online eye tracking tools could be used for follow-up work related to “Explainable AI” visualizations in future projects.

Anticipated Outputs

Blog posts; symposium presentation to Duke and RTP data visualization community; data for grant applications

Student Opportunities

Ideally, this project team will include 4 graduate students and 9 undergraduate students with interests or expertise in technical data visualization skills, cognitive or social sciences, and/or visual arts or design. 

Students will join one of three subteams focusing on testing different data visualization best practices. Each group will create relevant stimuli and collect eye-tracking, facial expression, questionnaire and qualitative interview data to assess audience members’ responses to the stimuli. Students will analyze and interpret the results.

The whole team will meet weekly to train in data visualization practices, discuss progress, organize weekly activities and request and give feedback. Each subgroup will meet separately each week to work on assigned tasks. 

All team members will develop skills in literature searching, running cognitive experiments, writing blog posts, presentation of findings, and evaluating, discussing and critiquing results. Graduate students will gain experience in mentorship and project management.

Timing

Fall 2024 – Spring 2025

• Fall 2024: Start training in data visualization and eye-tracking and expressional awareness software; design environmental stimuli, interviews and questionnaires; finalize experimental design; practice qualitative interviews; begin data collection
• Spring 2025: Run experiments with in-person and online participants; analyze data; generate blog posts; develop data visualizations; recruit visualization experts; present to local data visualization community

Crediting

Academic credit available for fall and spring semesters

Image: Data visualization of marine data, by Ars Electronica, licensed under CC BY-NC-ND 2.0