Rules Are Relevant
S-H-O-W conference speakers talk about maintaining the delicate balance between conveying meaning and exhibiting creativity
Amanda Cox, editor of The New York Times’s Upshot section, once said,
“There’s a strand of the dataviz world that argues that everything could be a bar chart. That’s possibly true, but also possibly a world without joy.”
Perhaps this is true of data visualisation and rules as well. If everyone uses the same rules, do we risk a joyless world of boring, homogeneous visualisations?
No! We all make different editorial choices and we apply the rules in different ways. Even if we were all given the same assignment, with the same rules, the outcomes would be unique. Even applying the same rules, we can all find different creative outcomes.
Rules are relevant, especially in data visualisation. Rules and conventions are needed to guide the maker to better understand the data and tell the story. Rules help the reader to better understand the information.
Rules in dataviz was the theme of the online S-H-O-W conference in November. Speakers and attendees explored the trade-offs between the clarity that rules provide for immediate understanding and stretching the rules creatively into something unique and attractive.
I asked a few of the speakers of the conference about their relationship with (designing) rules. Here’s what they told me:
If you had to pick one or two rules (or principles) that are important to you in your work, which one(s) would that be?
Andy Kirk (Visualising Data): Though there are three key principles I always maintain in my mind as I am making visualisation decisions, there is one that is first and fundamental — I must seek to make my work trustworthy. This assumes I am truthful, but goes beyond that — it is whether my audience believes what I have produced is reliable, accurate, and has integrity. Are there any inaccuracies that might undermine faith that I have made correct calculations? Are there any assumptions made that I need to express? Have I chosen to leave out certain data items that might not be apparent or clearly explained? Have I used any chart design methods that might distort or deceive the viewer into misreading what is portrayed? Though the judgment of trust may not always be the immediate thing crossing the mind of the viewer, it should be/will be a concern they eventually reach. You can have the most compelling, elegant, and fascinating visualisation in the world, but if it is unable to earn the trust of the audience, then it has fallen short of the required standards.
Katie Peek (science journalist and designer): I always make sure I understand the data and its limitations before I create the design. I always, ALWAYS speak with the person who created the dataset. (In my training as an astronomer, I learned all too well that every data set has its quirks and limitations, and the best way to communicate that nuance is through a phone or in-person conversation.) And second, I try to identify the central point I’m trying to communicate — I ask myself, “What is this graphic saying?” — and design the visualization to answer that question first.
Giovanni Magni (Accurat): The high-level principle that I would always keep in mind is to be accurate, to create visualisations where visual shapes are faithful to the original set of data, but this is not everything. When discussing rules in data visualisation, people usually refer to common practices used for the correct and formal representation of visual quantities but, from my experience, there is something more important than that. As a designer, it is necessary to be aware that every choice you make depends on the context. Even the strongest rule could be shaped and adapted to the environment where the final output would live, if this will make it more effective.
Sandra Rendgen (data visualisation consultant): My number one rule? User testing is king. Always test your visualisation work on potential users, be it on your mother, your aunt, your friends, or the store owner down your street. As much as you can try, it will always be difficult to avoid designing based on our own experiences and knowledge. Even the most experienced designer of visualisations will never be able to foresee exactly how people read and interpret a given piece of work.
Manuel Lima (VisualComplexity.com and Google): I have so many favorites that it’s hard to pick one or two. One rule that I see myself turning to often is the flexibility-usability tradeoff. It simply tells us that as the flexibility of a system/experience/product tends to increase, its usability tends to decrease. Finding the ideal balance between flexibility and usability is quite hard and arguably what sets apart great experiences from the rest.
Lisa Charlotte Rost (Datawrapper): For me, the key rule is, “The visualization needs to be clear about what it wants to achieve.” This principle has improved my visualizations in the past — or rather, related questions like:
- What’s the point of your visualization?
- What should your readers learn from it?
- What’s the statement you want to make with it?
- Why does your visualization exist?
- And being clear on that, how can you now improve it further?”
It’s not a good idea to use this principle for everything you do. There are lots of great data visualizations out there that don’t need to have a clear goal: data visualizations that are simply beautiful; data visualizations that let readers see connections that the creators didn’t think of; data visualizations you can get lost in for minutes and leave inspired. So often, it’s a question of background. I’ve worked in journalism and for Datawrapper — a data visualization tool that is widely used by people with a journalistic mindset. And in this bubble, the rule, “Be clear about what your visualization wants to achieve,” has helped me more often than not.
What are the biggest challenges in reaching or connecting with your audience?
Shadi El Hajj (RefractionLabs): We live in times of instant gratification and ever-shortening attention spans. Unfortunately, there is no shortcut to knowledge, and to date, the only way to convey nuanced information on a complex topic is through long form, be it written or visual. Data visualisation is synthetic by nature, and we therefore must be even more wary of oversimplification and amalgamation. In the same spirit, I try to avoid literality. Problems and experiences are complex, and I try to illustrate this complexity with graphics that aim to intrigue and raise questions, not give definitive answers. Even if this comes, sometimes, at the expense of clarity.
Andy Kirk: When you have a ‘known’ audience and can directly communicate with them, you are in a really fortunate position. You can have conversations where you establish where the intersection lies between what data and what content you have available to potentially communicate to them, and/or they can explain to you what it is they need. This two-way negotiation leads to a mutual understanding of what content will be most relevant. The biggest challenge, therefore, is when you don’t have this kind of access to dialogue with your potential audience. You have to make assumptions, using the best reasoning available to you, to anticipate what you think they would find relevant. When this audience is especially broad in the diversity of its characteristics, wants, and needs you can hit a point of inertia where the burden of choice becomes too large — there are too many mouths to feed. So, at some point, you have to determine a hierarchy of needs and find the cut-off point. You can’t satisfy everyone, but whose needs WILL you satisfy? That is a difficult editorial judgment to make.
Katie Peek: Because I work on the print page, I have very little control over the order in which someone reads the work. Maybe they’re starting with the key, maybe with the annotations. Maybe they never read the introduction! So I write each of those elements with the idea that they could be where the reader starts, and try to drop enough hints into the text that they’ll be able to get the main points no matter where they pop in.
Hannah Davis (generative musician): Above some level of complexity, a data sonification requires an audial ‘key’ to understand what is happening in the piece. This can be a little clumsier than having a visual key, since you have to internalize it and identify the different parts of the audio and remember what they mean in relation to the data. Because of this, I think it’s somewhat more interesting to use data sonification either, a) for data that you’re going to be interacting with over and over, like streams of data, where you can kind of intuitively learn the key over time, or b) for music creation, which is what I’m doing these days. Creating music doesn’t have the same issue of needing to learn the key and interpret the data — you can just use the data to generate more interesting outputs.
Boris Muller (Professor Interaction Design Potsdam): The biggest challenge is really understanding what questions the audience has. If the audience is not really interested in the data, they won’t be interested in the visualisation.
What is the biggest mistake often made in data visualisations when it comes to design rules?
Evanthia Dimara (assistant professor data visualisation Utrecht): A “mistake” that I often see and was partially my motivation for the theme of my S-H-O-W conference talk, is that although the visualization domain has made great strides in the development of semiology of graphical representations, it lacks rules on how to design for visualization operations. Compared to representation techniques, we pay much less attention to leveraging possibilities for novel interaction techniques and modalities. And by modalities, I refer to both human modalities (e.g., other senses besides vision) as well as technology modalities, such as other input devices beyond the desktop. Interaction has been overshadowed in visualization, and I think it is the key to place data visualizations in the hands of broader audiences.
Andy Kirk: Rules exist to give us reliable guidance when we most usefully need them, they keep us on the right track. They sometimes cause us to be complacent, though. Rules are often simple judgments that fall into the ‘always do this’ or ‘never do that’. Far too often in visualisation, things that are supposed to be considered as general guidelines get marked up — and preached to others — as rigid rules when they can’t be universally applied. Don’t label your axis? Might not need it. Don’t offer a colour legend? The colours may be immediately obvious in their association. Using red and green? Maybe your audience has no colour-blind members. What characterises the real world application of data visualisation is compromise and trade off — a pursuit of perfect is never possible, so how do you do your best? You need to be flexible. You need to demonstrate nuance in your approach so you can respond to contexual, editorial, or data challenges accordingly.
Katie Peek: Without speaking ill of anyone’s work (because we never know the circumstances under which it was made!), I will say that I believe in breaking rules if it will help the visualization serve the audience better. So perhaps the biggest mistake made with design rules is blindly following rules.
Giovanni Magni: The biggest mistake is to think that the rules are unbreakable. Or, to think that, if you follow those rules, you will automatically end up doing something good. Don’t get me wrong, you will probably make something formally correct, but this could be far away from an effective design. Too often the discussion revolves around the fact that a visualisation might be correct statistically, while too seldom we ask if the visualisation solves the task for which it has been created, that might be a totally different thing.
RJ Andrews (InfoWeTrust): Rules are not the end. They are merely a starting point.
Steve Haroz (research scientist Inria): Letting design rules limit how you communicate is an unfortunate mistake. As many so-called rules for visualization are little more than popular opinion with no empirical backing, they can cause many missed opportunities. One way to avoid this problem is to seek evidence to support a rule and to ensure that the evidence goes beyond a mere anecdote or the insistence of an “authority” in visualization. If you can’t find the evidence, the rule is just an opinion. And your opinion is as valuable as anyone else’s.
Manuel Lima: Design rules or principles are meant to be generative, not prescriptive. They are meant to inform and inspire the project, not provide a step-by-step guide.
Lisa Charlotte Rost: I don’t know. But I’d very much like to know. Please send an email if you know. Thanks!
The online S-H-O-W conference took place on November 27 & 28, 2020. The next conference will be organised in 2021. For more information about training sessions in dataviz, have a look at the website www.GraphicHunters.nl .
