A Middle Ground
We criticize flashy infographics and bad visualizations, but we also want to attract viewer's attention. We strive for accuracy and efficiency, but we also want to tell stories. We dislike chart junk, but we like beautiful charts. We need to find a middle ground.
Available Tools
We cannot give people visualization tools, or they will hurt themselves. There are too many ways in which naive users can create charts that are ugly and unreadable. Excel is the best example: millions of terrible charts are created every day, because people can. They are then shown on PowerPoint slides that do not give people any useful information. As a result, many people think that visualization is useless.
We have to give people visualization tools, so they can experiment, learn, and come up with interesting new things. Spreading the word about visualization is one thing, but it's not going to work if people can't try it on their own data. Not all of those visualizations are perfect, but it's the only way we will be able to demonstrate how useful visualization can be. Besides, people who haven't been indoctrinated by Tufte will be able to create new, exciting, and surprising things.
Beauty
Beauty of a chart only gets in the way. If a chart doesn't speak for itself and the data it represents, it's the wrong chart. No amount of dressing it up will make it any better. Shiny, pseudo-3D dashboards are what you get when you make your decisions based on beauty rather than usefulness, and we're not making fun of those for nothing. Beauty has its place, but not in visualization.
If we can't get people to look at a visualization, what's the point of creating it? Making it beautiful does not mean making it ineffective. Well-designed and beautiful charts are actually often more effective than ugly ones, and they make people pay attention. If a chart is ugly, nobody will look at it, and all the work that was put into it will be for nothing.
Chart Types
Picking the right chart means understanding the data and the task. It's not a question which chart you like best, it's about which one works for what you want to do. Just because people like pie charts doesn't make them the best choice for everything (quite the opposite, in fact). Plus, you can never go wrong with a bar chart.
Insisting on a small number of boring old chart types does not make for a very exciting visual vocabulary. If we want to take visualization further, we have to be able to experiment. Sometimes pie charts work, sometimes they don't. Sometimes areas work well for comparison, sometimes they don't. But if all we ever need are two or three different chart types with minor variations, what's the point of doing new things in visualization?
Complexity
Visualizing complex data means using complex visualizations. You can't dumb everything down to a single chart. We need complex tools for complex and high-dimensional data: scatterplot matrices, parallel coordinates, etc. If people want to understand the real world, they will need to learn to work with these tools. Either that, or depend on people to digest their data for them.
Complexity may be a virtue in analysis, but not in presentation. To effectively present data to a wide audience, we need to reduce the level of complexity and focus on the core finding or message. Getting people's attention is difficult enough, we can't scare them away again by bombarding them with data and complexity. If you need to make it complex, build it up slowly. Don't use your lack of understanding (or time) as an excuse for not boiling things down to their essence.
Infographics
Infographics are the bane of visualization. They are pretty and flashy, but they often contain horrible, meaningless charts. Their designers can't be bothered to read up on visualization, and the always use pie charts for everything. They waste a lot of space to tell you a few numbers that could be explained in a few sentences or a table. The only way to improve infographics is usually to throw them away and use a single bar chart instead.
Infographics can teach us a lot about getting people's attention and providing context. A well-designed infographic not only shows you data, it also creates connections between facts and explains. While visualization is good at creating general-purpose techniques that work for lots of data, infographics are specific and help people understand.
Telling Stories
A good chart tells a story, no matter how basic it is. In fact, the more basic and less cluttered, the more effectively the data can speak and tell its own story. There is no need to dress up a chart or create an infographic. Make it clear and simple, and people will spend the time to read and understand the data; the story tells itself.
Telling stories effectively is hard work and requires a lot of care. Many charts may work as the raw material for a story, but actually telling the story is a different matter. Telling a story means engaging the viewer, and providing context and a path through the data. A well-told story will make the viewer understand and remember the facts.
Conclusions
Visualization is still rapidly evolving and changing. Settling for what we have and insisting that it's the only way to do things will not help the cause. What we need to find is a way to use what we know while still being interesting. We need to be able to build on existing knowledge without being limited by it. We need to engage viewers and readers without selling out to flash and glitz. We need to find a middle ground.
Posted by Robert Kosara on May 10, 2011.