Dashboards: Review of Information Dashboard Design Second Edition
When I started working with data visualization, I read Information Dashboard Design; I did not get that much out of it. Some of the display methods Stephen Few was advocating were not easily supported by my tools. His theories were sound, but I never finished the book. When the second edition was released, I reread the book and found a new perspective. With a few years of designing under my belt – I had more respect and understanding for what he was advocating about display methods. Here’s a short book review and how I applied his theories to a dashboard plucked off the internet.
Learn to Design Dashboards
Stephen Few, Perceptual Edge, most likely is the world’s leading expert on dashboard design. He has written several books on data visualization and dashboard design including Information Dashboard Design, which was originally published around 2006.
In the book, he does a thorough job reviewing what is important about dashboards and why it’s important. In the second edition, he goes into more detail about sparklines (invented by Edward Tufte) and the bullet chart, which Few himself invented. The book includes some additional chapters on the analytics process and other helpful topics to young designers.
Many people get dashboards, portals, and reports confused. These are different ways to deliver information and people frequently say dashboard when they mean report available on the web. Early in the book Few noted his definition of an information dashboard:
“A dashboard is a visual display
of the most important information needed to achieve one or more objectives,
that has been consolidated and arranged on a single computer screen
so it can be monitored at a glance.”
This definition guides the discussion of the entire book. The part I see most people getting confused about is “single computer screen” meaning that it’s a succinct area. Another key point he makes is that the information can be monitored at a glance, which he defined as providing overview information and not requiring interactivity, such as filtering or brushing.
Dedicate 50% of Space to 3 Facts?
I swiped a random dashboard from the web – it’s only crime was that it showcased gauges as the main information delivery elements. It was easy to find because most data visualization vendors choose a gauge because it is a highly graphical element that captures attention. That’s the way advertising works, right? Car dealerships do not show you a plain gray minivan; No, it is a red convertible sits out front. If they can get your attention, then you can talk about my cars. Then we realize you need the gray minivan and I make a sale. We both know a 2-seater convertible doesn’t really match your lifestyle.
However, let’s think a little more about this example. It contains three gauges that explain with the artwork and text below it the values. At the top, you can use a drop-down menu to see each region. When implemented – it is really cool but over time don’t you think the user would simply read the text instead of looking at the artwork? If they want more details they have to drill down to the next level anyway.
Let’s not forget this is the most important real estate on the dashboard – the top half! You may not realize it from the image – but 50% of the dashboard was dedicated to three measurements: Revenue Growth, Profit Growth, and Profit Margin. It appears these are year-over-year measurements – so the user only reviews the dashboard yearly?
I’m not saying these are not important measurements but that is a lot of space for 3 facts even if it really shows it for all six regions. Seems like the user would want to know how the regions are doing when compared to each other, which is a bit of a challenge. You have to remember the Northwest was 14.5 and then click to Southwest to compare, and then go to the next measure while trying to recall all of these figures. Might be difficult to determine which region is doing the best and worse because you have to memorize it for six region? How is the company doing overall?
Ok – I’m being pretty rough on this humble dashboard most likely designed to show the product had a très chic gauge and no other thought given to dashboard design.
Check Out the Redesign
Few explained that dashboards need to impart as much information in as little space as possible. In the following figure, I have done a few things to display the same metrics but in a more clever way. I think it’s important to know how the company is performing overall, so that is the first part of the dashboard displays the same measurements but with some supporting information. Then without moving the mouse, you can compare all of the regions.
The manager can compare overall company and regional results. Look how the red jumps out from the soft colors – 2 regions are struggling and take a quick glance at the sparkline to see it’s just a continuing trend from the past year. Plus why is the profit margins so low in the other 2 regions – the expenses must be too high. [Note: I had to create the sample data so my trends and values are suspect.] My design is for quarter-over-quarter reporting but it could easily be updated for month-over-month. I imagine the region names linking to interactive reports that would take advantage of the SAS Visual Analytics capabilities and allow more user data exploration.
This design closely follows Few’s dashboard design theories and examples. He advocates soft colors, bullet charts, and sparklines. He doesn’t say not to use other data visualization methods. My purpose was to show you how much more information was placed in the same amount of space!
Ok – I agree it’s the gray minivan not the red convertible. This dashboard has its own sex appeal – you can get answers quickly and then explore other reports if you want more information. However, I’m wondering if this design actually works that well with SAS Visual Analytics tool – do you have an opinion?
Note: I did not receive any compensation for this review or the links contained therein.
Never miss a post!
Get the latest BI Notes post in your Inbox when a new post is released! Click here for free subscription. You email address is kept private and you can unsubscribe anytime. Go ahead ... join us!
Latest posts by Tricia Aanderud (see all)
- Can SAS Visual Analytics Replace PowerPoint for a Small Business? - November 11, 2013
- User Groups: SESUG in St Pete Beach was Great! - October 27, 2013
- Data Management: Who cares about Big Data - October 18, 2013
- Data Visualization: How to Avoid an Epic Fail - October 10, 2013
- Dashboards: Review of Information Dashboard Design Second Edition - September 23, 2013