How do you ensure that your data visualizations are accessible and inclusive for diverse audiences?

To help understand your audience without getting too personal, you could ask questions like: ?How do you like to engage with a data visualization? Are you more of a keyboard navigation person, mouse, or screen reader? This helps you understand how they engage which will also influence design choices. ?What is the extent of subject matter expertise for this area? The less expertise, the simpler you want to keep the dataviz so you're not overloading their short-term working memory (cognitive load). ?Do you find it helpful to have a dashboard annotated to highlight the key insights? This helps if someone is distracted and trying to do their work and if they have any cognitive challenges with understanding the data presented.

Asking open ended questions can be a good starting point. However, you will get more meaningful feedback when you share with the users an early version of the visualisation/dashboard or a prototype. This lets them see and understand what you are trying to build. You will learn a lot about their thought process just watching them use it and listening to their questions.

In order to understand your audience, it must be defined. If you are tasked to build an experience for more than one audience, you must identify each per visualization throughout that experience.

One thing to think about is the cultural significance of various colors in your design. For instance, the color red in Chinese culture symbolizes luck and happiness. Whereas in Eastern European countries, it is often associated with communism.

Color can be a powerful addition to your visualization. However, if your audience uses a screen reader, it's not powerful at all, because screen readers don't read color. This is why knowing your audience beforehand is so important. When you do use color, it's best to test the contrast against each other and the background. For data elements, the contrast ratio required is 3:1. For regular text the contrast required is 4.5:1 and for large text (over 18pt font), the contrast is 3:1. Use chart types that are easy to understand, especially weighing the audience's expertise (you don't want them to have to figure out the data and what a violin plot means).

A way to help make data visuals more accessible for persons using screen readers (without taking up valuable real estate) is to add a text explanation of the screen in the same color of the background so it cannot be seen. A screen reader will pick up the text, but the text will not be visible.

Providing access to the data and descriptions (like alt-text) allows not only screen readers to read the data, but allows people to access what's behind the visual. The tool you use may have a data window behind the visualization, which a person can access. You can also provide a link to the data at the bottom of your viz or in a help icon.

One way to test and evaluate your visualization is to use the free assessment, Chartability, which is focused on data visualizations. It can be found at https://chartability.fizz.studio. Getting the viz in front of your users and getting feedback during the development process is also really helpful so that you know what changes need to be made when it's pushed to production. Develop, evaluate, iterate, publish, evaluate, iterate, publish. This approach/mindset is helpful as needs and your audience change.

While tools and surveys are a good start, never underestimate how much you can learn from simply talking to users and observing how they interact with the end product. Observing and getting direct feedback about the user experience will bring obvious accessibility choices to the surface and can also lead to the creation of other insightful visuals.

I am of the opinion that visualizations should be kept as simple as possible, whilst maximizing the data presented to the viewer. Ultimately, this makes for a very clean, clear message regardless of who is viewing the visualization.

Here are some things to keep in mind: ?Design with accessibility in mind ?Keep your conventions consistent. Titles & headers mean something...don't just increase font size. Apply conventions—small typically means less…less size, less color. Keep your layouts consistent where possible so people know where to focus. ?Prioritize content. Most important item should be at top/left. ?Share how to interact with the visualization. ?Keep it simple and roll up the data, if possible (challenge, do we actually need row level data in this view?) ? Use white/blank space between elements. Giving the content room to breath on the canvas also aids in comprehension and avoids overwhelm. Give these tips and a go, and ask for help if you have questions!

Organise your dashboards as a narrative which follows a natural order, examples: high level metrics on top deep dives lower down, chronological order of occurrence, bottom line on top(most important info on top, supplemental info lower down) On dashboard documentation can be really useful. Define metrics, important terms and concepts used, etc. You can always go the extra mile and create a video tutorial on how to use it and add a link to the video on the dash.