BY RYAN BOWER – SR. DATA SCIENTIST
“THERE IS NO SUCH THING AS INFORMATION OVERLOAD, ONLY BAD DESIGN”
— EDWARD TUFTE
PROFESSOR EMERITUS OF POLITICAL SCIENCE, STATISTICS, AND COMPUTER SCIENCE AT YALE UNIVERSITY
This weekend, I went on a Lego-building binge with my son. Following his fifth birthday party, we had a lot of Lego work to do. Somewhere in the course of building submarines and spaceships, I realized that in all the Lego Universe, there are no words in the assembly instructions; you just learn to read the diagrams.
It is amazing that a whole world can be built using instructions without words. Here I was, a good twenty years after my last Lego assembly, and it was easy to jump right back in. It’s not simply phenomenal design work that enables this “Lego Literacy.” I think experience and understanding of the system play a role too.
Of course, this made me think of data visualization. I love visualizing data because to me, it’s far easier to understand data when it is displayed visually. I realize that is not the case for everyone, and bad charts are not always to blame. Kaiser Fung recommends a test for visualizations: if you can remove all text from a plot, and still understand the meaning, you have succeeded. I like this test, but I don’t think it tells the full story.
There are some plots that most viewers recognize. These are the types of charts that are available in Excel (for good reason!). Line-charts, bar charts, and even pie charts are all commonplace. If I show one of these types of charts in a meeting, I can be confident that nobody in the room will be confused. This represents the lowest level of visualization literacy. Done well, everyone can understand them, and there is no training required.
Despite the simplicity of these charts, there is power in their effective use. I have found that just highlighting a specific trend or “something to look for” draws users into a chart. Another way to empower simple visualizations is to build a dashboard, or standardized view of data. Not only is it easy to quickly see what’s going on, but it also becomes easier to spot changes. Another powerful way to use simple charts is to animate the charts to show changes over time.
Beyond the simplicity of standard chart types, there are visualizations that must be learned. Some of the most enjoyable work I’ve done has required building—and teaching others to use—a different style of visualization.
These charts are often unfamiliar to people—they require a level of visualization-literacy that most people do not have. That doesn’t mean we should avoid these plots. On the contrary, I have often found that the process of showing somebody a new way to view data enhances their understanding of it.
It excites me to think about visualization literacy. Developing this literacy can help people get to a better understanding of data. By visualizing data we often find useful pieces of information that cannot be found any other way.
But enough of visualization literacy, I need to figure out why this spaceship nosecone won’t fit onto the body.
Lego Visualizations
BY RYAN BOWER – SR. DATA SCIENTIST
This weekend, I went on a Lego-building binge with my son. Following his fifth birthday party, we had a lot of Lego work to do. Somewhere in the course of building submarines and spaceships, I realized that in all the Lego Universe, there are no words in the assembly instructions; you just learn to read the diagrams.
It is amazing that a whole world can be built using instructions without words. Here I was, a good twenty years after my last Lego assembly, and it was easy to jump right back in. It’s not simply phenomenal design work that enables this “Lego Literacy.” I think experience and understanding of the system play a role too.
Of course, this made me think of data visualization. I love visualizing data because to me, it’s far easier to understand data when it is displayed visually. I realize that is not the case for everyone, and bad charts are not always to blame. Kaiser Fung[1]Author of ‘Numbers Rule Your World’ and ‘Number Sense’. His blogs (ChartJunks and NumbersRuleYourWorld) are also highly recommended for those who love visualizations. recommends a test for visualizations: if you can remove all text from a plot, and still understand the meaning, you have succeeded. I like this test, but I don’t think it tells the full story.
There are some plots that most viewers recognize. These are the types of charts that are available in Excel (for good reason!). Line-charts, bar charts, and even pie charts[2]Most visualization experts recommend against pie charts. These charts are generally considered bad design because it is difficult to accurately compare sections of the chart. There are other ways to effectively communicate the same information, but pie-charts are nonetheless ubiquitous. are all commonplace. If I show one of these types of charts in a meeting, I can be confident that nobody in the room will be confused. This represents the lowest level of visualization literacy. Done well, everyone can understand them, and there is no training required.
Despite the simplicity of these charts, there is power in their effective use. I have found that just highlighting a specific trend or “something to look for” draws users into a chart. Another way to empower simple visualizations is to build a dashboard, or standardized view of data. Not only is it easy to quickly see what’s going on[3]See also, this previous piece about quickly recognizing what’s going on in a visualization or dashboard: https://elicitinsights.com/blog/think-fast/, but it also becomes easier to spot changes. Another powerful way to use simple charts is to animate the charts to show changes over time.
Beyond the simplicity of standard chart types, there are visualizations that must be learned. Some of the most enjoyable work I’ve done has required building—and teaching others to use—a different style of visualization.
These charts are often unfamiliar to people—they require a level of visualization-literacy that most people do not have. That doesn’t mean we should avoid these plots. On the contrary, I have often found that the process of showing somebody a new way to view data enhances their understanding of it.
It excites me to think about visualization literacy. Developing this literacy can help people get to a better understanding of data. By visualizing data we often find useful pieces of information that cannot be found any other way.
But enough of visualization literacy, I need to figure out why this spaceship nosecone won’t fit onto the body.
Footnotes [ + ]
Tags: