37th Annual CSUN Assistive Technology Conference Has Concluded

Making Graphics and Charts Accessible: Case of MATLAB Online

Date & Time

Tuesday, March 15, 2022 - 3:20 PM PST

Location

Grand GHJK

Description

Charts are everywhere. They are widely used in classrooms and business settings to communicate information in a digestible format quickly. Plotting and visualizing data is an enabling step for further analysis in many scientific and engineering workflows. In introductory technical courses, generating charts and graphs from data is one of the first activities tried by novice users of technical computing software. In recent times, the ready availability of rich datasets has also resulted in the broad adoption of data-driven journalism, where charts, maps, graphs, and other visualizations are used as a narrative tool to tell stories. This has resulted in countless charting apps and software products tailored for specific workflows. Yet, despite this explosion of new charting and visualization tools, the accessibility of graphics remains an open challenge. In particular, there is a need to accommodate accessible ways for consistently navigating the graphical outputs within the context of navigating the complex software environments. Additionally, visual elements in charts and graphs need to be made accessible via the screen-reader to allow visually-impaired users to consume the information presented in them. Challenges also arise while providing accessible data exploration tools and interactive workflows like panning, zoom, and rotating actions. Existing standards, such as WCAG 2.1, provide limited guidance on making data visualizations accessible. Additionally, charts and visualizations come in various types with diverse underlying structures. Because of this lack of a uniform structure, the hierarchy of the elements is not always straightforward. A rich visualization could include context-setting elements like titles, axis labels, tick marks, and legends. Additionally, the content may be presented using a combination of lines, bars, markers, and text. Furthermore, annotations like arrows and circles can provide additional commentary on the plotted data. Incorporating accessibility principles in the graphics space requires careful consideration of this unique set of challenges. MATLAB Online, the web version of MATLAB developed by MathWorks, is a scientific computing tool widely used in academia and higher education by scientists and engineers. Currently, MathWorks is improving the accessibility of MATLAB Online, starting with keyboard and screen reader support. Scientists and engineers with limited motor skills, who find it challenging to use the mouse, rely on keyboard-based navigation and interactions for graphics workflows. Users with vision difficulties, who cannot visually spot trends or points of interest depicted in charts, rely on screen readers as an output mechanism. But because of the challenges mentioned above, extending keyboard and screen reader accessibility to graphics within MATLAB Online presented many interesting design challenges. The solutions we will illustrate in our presentation cover the approach we adopted to incorporate hierarchal navigation of charts and graphs. Our approach operates at two levels. At the first level, figures, which can contain multiple graphs and charts, are accessible via the keyboard and screen reader. The announcement for the figure gives a high-level descriptive summary of its contents and orients the user for further navigation. At the second level, keyboard access is extended to individual charts, which are coupled with screen reader announcements. We first allow focus to fall on context-setting elements (like titles and axis labels) and then on to the chart’s contents such as lines, bars, scatter markers, texts, patches, etc.

Audience

Higher Education
Information & Communications Technology
Government
Media & Publishing
Research & Development

Audience Level

Intermediate

Session Summary (Abstract)

Charts contain various context-setting, stylistic, and graphics elements that lack uniform structure, presenting a unique set of accessibility challenges. We discuss improving keyboard and screen reader accessibility in charts by taking the example of MATLAB Online, a complex single-page web application where graphics are rendered within an HTML canvas element.

Session Type

General Track

Topics