Shiya Cao is a MassMutual assistant professor in statistical and data sciences at Smith College. Her research centers on disability inclusion and broader social inclusion topics using quantitative, qualitative, and design science methods. She is eager to help the statistics community pay attention to human, social, and emotional elements of data science.
People who are blind and visually impaired represent a substantial segment of the population, with 4% of 54,204 students in the 2022 American College Health Association survey reporting they are blind or have low vision. Those students frequently do not have access to data visualizations taught and used in postsecondary statistics and data science classes. Making data visualizations accessible to blind and visually impaired people would help improve equity in higher education and assist them with data-driven reasoning and communication.
The Guidelines for Assessment and Instruction in Statistics Education (GAISE) College Report lists creating and interpreting graphical displays as one of the nine central goals for introductory statistics. As statistics and data science educators, researchers, or practitioners, we should practice how to design accessible data visualizations and use them in a way that is inclusive of the blind and visually impaired community. Following are steps to achieve this goal.
Provide a Textual/Audio Description of the Graphic
According to the ACM Digital Library article “Visual Cues for Data Analysis Features Amplify Challenges for Blind Spreadsheet Users” by Minoli Perera and coauthors, a key question is how to present visualizations nonvisually. Providing a textual/audio description of the graphic is a common practice and has been well studied.
The best practice is to provide an alt text description for the graphic so blind and visually impaired students can use screen readers to comprehend the textual information and thus the graphic. However, alt text descriptions are often not informative or expressive. Creating a description that is concise and emphasizes the story using visual metaphors is key. Then, if needed, authors can point to longer descriptions of the graphic elsewhere.
For instructors who use R, the BrailleR package can make graphics with automatically generated screen reader descriptions.
Graphics: Be Specific with Words
In addition to using screen readers to convert textual descriptions to audio explanations, instructors can provide detailed audio descriptions in person or by recording. When doing so, the best practice, according to Rachel Friedensen and coauthors in “A Systematic Review of Research on Faculty with Disabilities,” found on EBSCOhost, is to be specific with the words spoken to discuss the graphic. This approach can help all students better interpret graphs.
Use Screen Reader Tools for Data Visualizations
Screen reader tools such as SAS Graphics Accelerator and Apple’s VoiceOver can support blind and visually impaired students when they are exploring graphics, according to Jonathan Zong and coauthors in the Computer Graphics Forum article “Rich Screen Reader Experiences for Accessible Data Visualization.” These tools can provide textual descriptions of a graphic, such as its encodings, axis labels, and ranges. They can also turn the graphic into an easy-to-read table (SAS Graphics Accelerator) or report descriptive statistics (Apple’s VoiceOver), as well as play a sonic representation of the graphic.
Make Data Visualizations More Accessible
Having visual guidelines that make graphics more accessible is helpful. For example, use high-contrast and color-blind safe colors, do not rely on color when differentiating data, use thick, bold lines, use a font that is at least six points, avoid highly ornamental fonts, minimize the use of italics and underlining, and consider zoom compatibility.
Create a Tactile Graphic
Blind and visually impaired students may want interaction with graphics that moves beyond textual or audio descriptions. The GAISE College Report also highlights the value of physical exploration to facilitate active learning. Creating a tactile graphic can help with this goal. One inexpensive hands-on approach is using handy materials such as wax string, dominoes, push pins and hot glue guns to “draw” graphs with raised lines, bars, or edges, so students can touch, create, or modify graphs.
Using the BrailleR package can turn a graphic into a scalable vector graphic or other similar file type for optimizing printing graphics via an embosser. Instructors can discuss with the university’s accessibility resource centers the materials and equipment they have to create tactile graphics.
Consider a Multifaceted Approach
When teaching data visualizations for blind and visually impaired students, I recommend using a multifaceted approach. For example, create tactile graphics and clearly and explicitly explain those graphics when students are touching and feeling them. Instructors can also make the data visualizations more accessible by recording the audio explanations or creating textual descriptions so students can review them after class.
Leave a Reply