Part of the DataVizLive interview series, exploring the conference sessions deeper, as well as showcasing other leading lights in Data Visualisation who are helping to improve how we use and present data for commercial success and social good.
Social Ambassador, Tableau
Advisory Board Member
Data Literacy Project
Colouring with Data
Allen Hillery talks with Alan Wilson, Principal Designer at Adobe, and one of the driving forces behind many of their industry-standard style guides and interfaces. Coming to the world of data via a career start in graphic design, Alan brings an aesthetic perspective to the conversation on how to visualise data science and user interfaces.
Alan will be presenting this session at DataVizLive on 10 November – sign up here to watch live and ask questions directly, and also get on-demand access after the event.
Ask Alan Wilson what his biggest career endeavour is, and he will most likely tell you that it’s to make data easier to understand through data visualisation. Alan is a Principal UX Designer at Adobe, where he describes his role as “working with his team to keep the Adobe Marketing Cloud looking good.” While his extensive background in graphic design qualifies him for this task, he aspires to ‘colour outside the lines’, so to speak.
Unlike many peers in his field, Alan has not shied away from becoming more data literate. He’s not daunted by maths and statistics, and he emphasises form and function in his design work. What I enjoyed most about my conversation with Alan is his unique path to data visualisation. Most people in the data visualisation space enter with a data background grounded in maths and statistics, and then pick up their visualisation techniques along the way to communicate data. Alan, however, is grounded in colour theory and sees design as a tool to make data useful. I was very eager to talk with him about Colouring with Data and how he goes beyond the aesthetics.
Allen Hillery: I loved your talk on Colouring with Data. You cover a lot of great topics on the use of colour and some key guidelines and best practices. Can you tell us about your career journey and what’s led you to this point?
Alan Wilson: I began my professional career as a traditional graphic designer with a degree in Graphic Design from Brigham Young University (class of 2004). In the beginning, I did a lot of branding and packaging design, but by 2007 I was doing mostly web design. I joined Adobe about a year after they acquired Omniture and began creating (what we now call) Adobe Experience Cloud.
At Adobe, I created Style Guides and other resources to help coordinate the efforts of our growing design team. These evolved to become design systems – predecessors to our design system today: Spectrum. I was also very keen to improve the display of data in our products. With the help of some really smart engineers, we developed a framework on top of D3 that vastly improved the quality, performance, and usability of our charts.
This is also the project that sent me down the rabbit hole of visualisation design—nearly 10 years ago. Today I split my time between data visualisation work and user experience best practices.
AH: As you open the discussion, you share how colour impacts everything from our sense of depth, shopping for produce, to how we chase leaves in the fall. Can you expand on this thought for the audience?
AW: As humans, our brains process a tremendous amount of visual information. The raw signal that our eyes send to our brains is not unlike a computer screen – a grid of colours. From this rather limited set of information we see and understand the world in tremendous detail. By reverse-engineering this we can make our own creations more intuitive, useful, and (hopefully) beautiful.
AH: I love how you walk us through the science of colour but always tying it back to the goal of “Does this chart answer the question?” One thought I had during your talk was: should we approach our sequential colour scales differently when designing for discrete vs continuous measures?
AW: I always approach sequential colour scales as a gradient – a continuous measure – to start. At some point I may make them discrete to make the visual easier to understand or because it’s a better reflection of the raw data.
AH: Is there a guide you use when deciding a sequential vs diverging colour scale?
AW: When deciding between sequential and diverging scales, I always look for a meaningful middle point. Sometimes it’s super obvious, like it is in many sports. A tie in the score is meaningful in the middle, so a diverging scale makes the most sense. In others, it’s less obvious, and more subjective like it is when looking at test scores.
Is there a useful middle point for the average score, or would it be better to represent it as one continuous range? In these cases, I try to let the narrative decide for me – I use the colour scale that will better communicate the information.
AH: Does someone in your line of work approach everything in life with colour?
AW: My kids sure like colour. And they’re very generous with it too. I often find splashes of colour on my walls, in the car, and even the floor!
AH: When do you feel it’s appropriate to design in greyscale?
AW: There are times when you need to iterate and work through design problems without the burden of details. Sometimes colour is a detail that is best left for later, after more fundamental problems have been worked out.
AH: This is similar advice I heard in Bridget Cogley’s talk on Busting Colour Myths. She recommended designing a dashboard in greyscale to take advantage of other design features, to accentuate a detail like font size for example.
How would you describe the ideal scenario where colour is used to highlight a key point or trend in the data? For instance, using bar charts or line graphs that are all grey except for the key month or category that uses a colour to make a statement. Where does this fall in your design approach?
AW: I really like this technique. I especially like it when using small multiples. You have a grid of charts – let’s pretend they’re countries. Each chart has the data for ALL the countries, but only one is coloured. The others are greyed out and fades back to provide context. But there are lots of situations where colour is used to create emphasis. I like to use colour on interactive charts to make it clear what is being interacted with.
AH: I love that technique as well. You talk about two visualisations that use colour very skilfully during your talk. The Kickstarter Visualization by James Wenzel and Matt Daniels uses an array of colours. The visualisation works as the user interacts with it to filter projects by category and city.
The Immigration Flow Through Tree Rings by Pedro Cruz, John Wihbey and team uses colour and position to explore the flow of immigrants into America over the decades. Can you advise the audience on finding the middle ground of being mindful of ‘No Rainbows’, while encouraging ways to push the boundaries and make it work?
AW: The ‘no rainbows’ guidance is firmly rooted to sequential colour palettes, where colour is representing numbers. Rainbows have a ton of shortcomings when it comes to this task. In both these examples, colours are being used for categorical encoding where a rainbow is representing dimensions, and a rainbow works brilliantly for that task.
AH: You’ve had a career journey that you’ve described as being super rewarding and empowering. Any final piece of advice for the audience?
AW: I thought maths was boring until it became visual. Data visualisation has layers of utility. Constantly ask yourself, “Does this chart answer the question?”. When it comes to guidelines and best practices – rules were made to be broken. Design is a tool to make data useful. Don’t be afraid to break the rules.
I found Alan’s enthusiasm for colour and its effective application in data visualisation to be infectious. I find his journey to data visualisation equally interesting. Alan’s always had an interest in numbers and was able to find a way to use his design skill set to be an advocate for data literacy. It’s this attitude that will bring organisations closer to achieving data democratisation. I recommend watching ‘Colouring with Data‘ on 10 November to all those in the data visualisation space.