Log in to like this post! How to Create Effective Dashboards – More Q&A Tobias Komischke / Monday, January 26, 2015 We had a great turnout for the webinar on Dashboard design, thank you for joining! There were so many great questions and comments about User Experience and Design that we could not cover them all during Q&A. Because of this, I’ll address some more here in this blog. How to cope with a lot of data and concepts to visualise but can't fit in one screen? The first consideration should be: what information is critical and needs to be on the dashboard and what information can we leave off? Oftentimes we tend to put too much on a dashboard and because of that we run into issues with screen real estate. After that, you can make use of tactics like these (and they’re covered in the webinar): (a) Use visualization types that are optimized for small screen real estate, for example sparklines and bullet graphs (b) Use graceful degradation of details where you remove non-essential details and focus on the core information you want to show (c) Use the same real estate to provide alternative views on the same information, for example upon a swipe replace a chart with the underlying data table (see screenshot above) And then, you can also provide the option to drill into more details, for example by viewing a chart in full-screen mode or providing an extra “Details” tab. The prerequisite of doing this is that the parent dashboard itself already provides sufficient overview and insights so that drilling into more details is truly optional. What program are you using for the swiping capability? This is our reporting tool ReportPlus, available for iOS and Android. What level of user is your software designed for (as a creator of dashboards)? e.g. User/Super User/Specifically trained to use it or basically what do I need to know before attempting my first venture into dashboards? We try to address all levels of users for ReportPlus. You don’t need to get trained on it. Even if you don’t know this type of tools, you should be able to use it. For example, even though a casual report creator may not know the concept of data connections and all the options that come with it, s/he can still pull data from an Excel document and turn that into meaningful information. What's your opinion on customizable dashboards, in which all users can pick and choose components to have, versus a static board of information provided by management? Personalization in and of itself is a good thing. It also helps with adoption of a dashboard, because people can make it their own. But then, if everybody is looking at their own information selectively, it defeats one of the main purposes of dashboards which is sharing the same information that is mapped to the same goals and strategies. Therefore, if I created a dashboard for managers that portrays information relevant to all of them, I would not allow them to change WHAT the dashboard shows but only HOW things are shown. For example, they should not be able to rip out one chart that shows year-to-date revenue (or whatever is tracked in this company), but they can change the color theme and the layout of the dashboard. In regards to the worst practice of "Exceeding boundaries of single screen" I'm currently working on creating a Dashboard for my team to consume and currently have two distinct categories I'd like to show. General Usage and Debugging visualizations. What is the best way to approach showing these two categories without invoking tabs? You should evaluate the cost/benefit for (a)having both on one screen for fast overview yet without as many details vs. (b) separating them on two screens/tabs so you’ll never see both categories in parallel but for each one you can display more details. How important is it for your team to see both together? How many details are enough to show to them? To me, this is the primary question. If parallelism is really not that important, then finding a design solution may include tabs, vertical scrolling or separate browser windows (if the users have dual monitors, they would get parallelism back that way!). What is your experience using Maps (GIS Info) related to a chart ? Is this a good practice ? Sure, that’s a standard part of a dashboard. The trick there is not to go overboard and for example embed Google Maps without any constraints so that users can pan way outside the pre-determined geographic regions that you want to report on. Also, the level of details on the map should not be too high because that would distract from the information you want to show. Refer to this video that shows how we went about building a dashboard. This dashboard has a map in it. You’ll see it’s a pretty simple map – not meant to allow you to find out the shortest way from your house to the next restaurant, but only to put relevant data into a geographic context. Traffic Lights (with different shapes) are an excellent way to expose data; you should recommend them like sparklines. Yes, the traffic light metaphor is good, because it’s universally understood. Of course, the limitation is that you can only code 3 gradations with it. I like that you mentioned the need for different shapes! That’s a good practice to accommodate for color blindness issues where people may not be able to differentiate red from green on the traffic light. So by encoding your values additionally with shapes (or symbols, sizes, etc.) you don’t rely on color alone. How to optimize the use of colors? Makes sure that… … the colors portray the right meaning (red is bad, green is good, etc.) … you don’t use too many colors. Beyond around 7 to 8 hues, it’s hard to tell them apart … the colors are different enough in hue/saturation/brightness so that you can tell them apart even when they’re shown close to each other … you have enough contrast not only between your chart colors, but also between chart colors and the background color of the canvas/the dashboard … the sum of your dashboard items look like one harmonious dashboard … you test your design with users! See also the answer to the previous question. How do you decide which data to display in the dashboard? How do you do your user research? That’s a great question that is hard to answer shortly. The executive summary would be like this: don’t design in isolation without ever leaving your desk. Engage with stakeholders and users. Rather than asking them about how you should design a dashboard, learn from them how they work, what is important to them, what information needs they have, what success data/metrics their company has agreed upon. Refer to this video that shows how we went about building a dashboard. It tells the whole story about how we researched the need for a marketing dashboard. What is your opinion on animating dashboards? Do they enhance the usability or cause distractions? If animation is offered as an option, I think it’s great because it makes the dashboard more engaging and may reveal more insights. I would not animate things per default, that would just make things too busy and distracting. For a really strong example about the power of animated data, refer to this sample we built. Here’s the link to the dashboard design video again. See you next time!