Moritz Stefaner works as a “truth and beauty operator” on the crossroads of data visualization, information aesthetics and user interface design. With a background in cognitive science and interface design his work balances analytical and aesthetic aspects in mapping abstract and complex phenomena. Moritz:”it’s a classical good news / bad news situation: Data visualization is easy to learn, but hard to master”.
Graphic Hunters: I have read that someone has to understand a visualization within seconds, others say visualizations should leave something to explore. What is your opinion on this?
Moritz Stefaner: Humans are very visually oriented, and in fact, the brain starts at the back of our eyes, with the first layer of neurons processing the incoming information. Data visualization has the unique power to bring phenomena outside our sensory reach into the experientiable realm — and make it accessible to this extremely rich and quick channel.
Great data visualizations are designed to immediately graspable, and readable without conscious effort within milliseconds. Ideally, you don’t even notice looking at a chart, but seem to look through it, at the underlying phenomenon of interest. So, this is the “fast track” access to information great data visuals can enable.
At the same time, great data visualizations always invite explore further and study deeper, after the key points of interest have been identified. What is the texture of the data, what are the distributions? What other things can we learn? How can we explain the interesting patterns?
So, the key is really to draw attention quickly and draw people into the graphic, but to keep them going with sustained interest to explore context and causalities.
I like to say: good data visualizations tell a story, but really great ones tell a thousand stories — but not all at once.
What is the first most important question one should ask before starting visualizing the data?
For data exploration, it can be exciting to just look at a dataset without preconceived questions to answer, and just explore what it can tell us.
When it comes to communication, however, the main question is really — why anyone would care about this data? What makes it interesting, unique? What’s the point, ultimately, of publishing this data set? Too often, we fall into the trap of publishing charts merely to show how much data we have collected or how complex our domain is. Also, being an expert in a domain, we often tend to forget how a novice might look at the same piece of information.
Great chart design comes from a position of clear understanding of what is most relevant to communicate, and how we can express it most effectively in visual language. So, understanding your own, but also your audience’s objectives and needs is really key to designing a successful visual.
What is the role of design in a visualization?
Often, people think design is mostly concerned with the “look and feel”. While these are certainly important aspects of user experience, good design starts at a much more fundamental level. As Steve Jobs famously said — “Design is not just how it looks, but how it works”. So, ultimately, design is concerned if the artifacts we put out into the world have the intended, positive effect.
So, thinking of scientific communication: from the title of a paper, over the metaphors or analogies you use, to typography or the color scale of a figure — all these are components in an overall approach to communicating your research and ideas. And the more you see this as a complete “design package”, the more effective your communications will be – both to expert and public audiences.
What lessons do you want the participants to take home for the training?
Well, it’s a classical good news / bad news situation: Data visualization is easy to learn, but hard to master.
Focusing on the first part, I will provide participants with a crash course in visual language and clarity in chart design. There are a handful of really useful facts and techniques that will help you to avoid classic pitfalls and produce visually clean and tidy charts. This stuff is really learnable – you don’t need to be an artist to create great-looking, effective figures, and a few techniques go a long way here!
After the basics are covered, I will show you techniques how to approach complex design problems and analyze the works of others in a productive way, speaking from an experience of almost 15 years of designing and producing data visualizations.
We’ll also discuss more advanced topics such as storytelling and narrative techniques, the role of text and annotation, and of course, if rainbows colorscales are evil indeed.
Overall, I provide participants with a design perspective and a new mindset to move forward in their own personal journey towards more effective visuals, rather than teaching software tools.
Also, I like to do lots of exercises, because in my experience data visualizaton is best learned by doing, then reflecting and discussing it together. It’s always a lot of fun to see what the group brings to the table and which solutions they can come up with for my challenges.