Making interactive data visualization accessible

My employer deliberately published no org chart. I built accessible radial prototypes so people could see how the business fits together. Leadership accepted the sunburst approach; implementation still depends on an agreed data owner and privacy requirements.

Problem
The company deliberately published no org chart, so people couldn't see how the business was organized or who to involve in a decision.
My role
Accessibility Program Manager
Result
Leadership accepted the sunburst approach.
Status
An accepted prototype direction, awaiting decisions on data ownership and privacy.
Organization
W.W. Norton & Company
Type
Inclusive design
Areas
Accessible data visualization, Accessibility

At a glance

Problem and constraints

As an accessibility program manager, my job was to make sure accessibility was part of how the whole business worked: design and engineering, sales and marketing, editorial, digital production, legal and customer support. Doing that meant knowing who did what, and who to bring into a decision.

The company had no org chart. That was deliberate: the policy was not to publish one, so as not to overemphasize hierarchy. I had never worked anywhere without one.

When I talked to colleagues, many felt the same gap. A conversation with a colleague in the company’s diversity, equity, inclusion and accessibility (DEIA) employee group changed how I saw it. Without an org chart, people joining the company had no way to see the social and political dynamics that shape how decisions get made. A policy meant to play down hierarchy was making the existing structure harder for newcomers to read.

So the problem was more than missing information:

Decisions and implementation

  1. Listened to colleagues across the business about what they needed to understand, and why.
  2. Built on a colleague’s proposal to leadership: a spiral, as an alternative to a top-down chart.
  3. Prototyped two radial designs in D3.js: a sunburst, and a collapsible dendrogram (a tree diagram you can expand and collapse).
  4. Designed for progressive disclosure, so people see a few levels at a time and choose where to go deeper.
  5. Designed for flexibility: several ways into the same information, so people choose the path that suits them.

How the work ran

This was a side project: a passion project I took on at Norton in 2025, alongside my role. It wasn’t part of anyone’s plan, so the prototypes had to make the case themselves.

I shared them with several colleagues first, then presented them to the DEIA employee group, and then to C-suite executives. By then I was showing two directions side by side: the sunburst, and the collapsible dendrogram.

A circle instead of a ladder

A colleague had already proposed a spiral to leadership as an alternative to a conventional org chart, and it seemed promising. I brought my background in data visualization and the technical side of accessibility to it, and built two prototypes.

A radial layout starts from the center and works outward. Teams are slices, and the size of each slice shows how many people are in that part of the organization. Nobody sits at the top of a page. The structure is still there, but it reads as how the company fits together rather than who ranks above whom.

The sunburst prototype for a fictional company. In the sidebar, the search box contains 'senior editor' and reads '4 people match'. Below it, the Browse the organization list shows the path from the chief executive to the Editor-in-Chief, whose four senior editors are highlighted as matches, each with the size of their team. On the right, the ring chart is zoomed into the Editor-in-Chief's department, with the senior editors in the inner ring and their team members in the outer ring. A panel below the chart names the Editor-in-Chief.
The sunburst, searching for senior editors and zoomed into their department. All names are synthetic.

Try the prototype. It opens a full-page demo built on synthetic data.

Showing less, on purpose

People liked both the sunburst and the dendrogram, but the sunburst’s visual simplicity and elegance won out. It also handles complexity well. Instead of showing every person at once, it shows a few rings from wherever you are. Choosing a team zooms in so that team fills the circle, and choosing the center zooms back out.

That keeps the amount on screen manageable. People can follow one branch of the company at a time instead of scanning a wall of names, which lowers the cognitive load of a large organization. It also puts the reader in control of how deep to go. Doug Schepers writes about designing data for cognitive load in the Urban Institute’s accessibility guide to data visualization (see Further reading).

Flexibility as an accessibility principle

One idea ran through the whole project: flexibility is an accessibility principle. People take in information in different ways, so a visualization shouldn’t prescribe a single path through it. It should give people autonomy in how they interact with the data, and the design has to build that choice in from the start.

The org chart offers several paths to the same information:

The paths stay in step. Opening a team in the list zooms the chart, and clicking the chart opens the team in the list, so people can switch between them whenever they like.

Making it work without a mouse

An org chart is no use if it only works for people who can see it and use a mouse, so accessibility was part of the design from the start. The sunburst had a menu that jumped straight to a person, and the same structure was available as a list.

My first versions made every person in the chart a separate Tab stop: 152 in the sunburst. That met the letter of keyboard access but not its spirit, because a keyboard user would have to press Tab up to 152 times to get past the chart. I reworked it:

I tested it myself with a screen reader and with a keyboard alone, alongside automated checks with axe. That’s a first pass, not proof that it works for the people who depend on it.

Results and limitations

Reflection

Some constraints are philosophical, not technical. Building an accessible chart was the easy part. The real constraint was a deliberate policy. The policy reflected a concern about emphasizing hierarchy, so the design needed to address that concern. Rather than argue against the principle, it was more useful to look for a creative answer that respected it, like a circle instead of a ladder.

Showing a working prototype made the discussions more concrete. A working prototype lets people grasp an idea instead of imagining it. When stakeholders can see and use the design, they can react to something real and iterate on it with you, instead of judging a proposal from a distance.

Further reading