In a world where decisions of all kinds are based on information derived from large datasets, ensuring people have access to information and understand its implications is more important than ever. The way most statistical information is made accessible is visually using charts and graphs, and the choice of which visualization to use is generally guided by the nature of the data to be communicated. However, research led by ATLAS PhD student Keke Wu finds that for those with intellectual or developmental disabilities (IDD), some kinds of data visualizations are harder to interpret than others.
“Keke's work is really the first in the visualization community to explore the idea of cognitive accessibility,” says Wu’s advisor, Danielle Szafir, an assistant professor of computer science with ATLAS and director of the institute’s VisuaLab. “It was previously an invisible disability to the community; we were completely unaware that common best practices were creating barriers for a large number of people.”
For this pioneering work, Wu and her co-authors earned a Best Paper award from the 2021 ACM CHI Conference on Human Factors in Computing Systems, the premier international conference on Human-Computer Interaction, which took place this week.
Approximately one in six children in the US has one or more developmental disabilities or other developmental delays, according to the Centers for Disease Control and Prevention. People with intellectual and developmental disabilities (IDD) may struggle with abstract thinking and spatial reasoning and, historically, have had limited exposure to mathematical and statistical training at school.
But despite the large number of people with IDD, visualizations such as pie charts and line graphs are typically provided without consideration for the population with whom they wish to communicate. As a result, those with IDD may struggle to make sense of some kinds of visualized data, says Wu.
The project is a collaboration with CU’s Coleman Institute for Cognitive Disabilities, which is particularly interested in how visualizations can support people with IDD in decision-making around financial self-advocacy. In addition to Szafir, Wu's co-authors on the paper include Emma Petersen, who graduated this spring from the ATLAS Creative Technology and Design master’s program; Tahmina Ahmad, an undergraduate majoring in computer science; David Burlinson, a post-doc; and Shae Tanis, co-director of the Coleman Institute and on the faculty of the CU Anschutz Medical Campus.
Data and social equality
At the start of her research, Wu identified three visualization design elements that could improve data accessibility: chart type, chart embellishment and data continuity.
Together, the team conducted a web-based mixed-methods experiment with 34 participants with and without IDD over Zoom. They gave data visualization tests to both populations, measured their test completion times and accuracy, and conducted interviews with the participants about the strategies they use to make sense of data. Wu then summarized the findings into four accessible visualization design guidelines.
The researchers concluded that the best chart type differs for people with and without IDD and that pie charts should be avoided for those with IDD; that discrete data representations, such as using isotype visualizations–where discrete objects are stacked together– instead of bar graphs will lead to more accurate performance for people with IDD; that semantically meaningful chart embellishments (such as using dollar signs with charts with financial information or stick figures to represent people) will enhance data interpretation for people with IDD; and that the visual complexity of the visualization would need to be managed.
Wu says future research will include a remote participatory design workshop “to see through the eyes of people with IDD” so as to better understand how they approach data.
“It’s not just to come up with hypotheses or just do a test with people, but really engage them in this process,” Wu says.
And she says her research may benefit others, even those without IDD. “Data can be intimidating even to people without disabilities. Our project starts with a particular population and has the potential to go to a broader one.”
About their Best Paper award, Wu says, “CHI confirmed our vision and the potential broader impact of our work, not just for a particular population, but for social equality.”
An imaginative journey to informatics
Wu’s personal journey to the field of informatics is an interesting one. As a documentary filmmaking student in China, Wu encouraged others to tell their personal stories in front of a camera. Later, as an exchange student in Maryland, she designed an app to encourage a severely depressed friend, using skills learned in a multimedia design class, and then wrote a workshop paper about it.
“I care about people,” says Wu. “And I care about society.”
Realizing that technology could be a gateway to helping others, Wu applied to ATLAS Institute’s MS-Creative Technology and Design (CTD) program to hone her technical and user-centered design skills. Soon after joining ATLAS, she met Danielle Szafir and joined Szafir’s VisuaLab, where Wu pivoted her academic focus to researching how people with IDD perceive data visualizations. She subsequently joined the ATLAS doctoral program.
“When I came to ATLAS, I didn’t have a very heavy technology background,” said Wu, who was also recognized this year as an Adobe Research Fellowship finalist. “This project and Danielle led me to become a researcher.”
