The best dashboards are informative, clear, and beautiful. You can make operational decisions by quickly glancing at them, or at most performing some light analysis. Stephen Few, a leading information technology teacher and consultant who focuses on best practices for creating visualizations of business data, has said:
"Operations dashboards must be designed to support real time situation awareness. They must grab your attention when it's needed, they must make it easy to spot what's most important in a screen full of data, and they must give you the means to understand what's happening and respond without delay. To do this, they require expert visual design and must express measures of performance clearly, accurately, directly, and without distraction."
Here are some best practices for making dashboards.
- Determine your audience and why it needs a dashboard: First, determine who the audience of your dashboard is, and which questions they need to answer by using your dashboard. The answers to these questions should inform every design decision you make. This information is especially useful when deciding which information is relevant and should be included in the dashboard, and which information is not. When possible it's best to make a dashboard for one audience, so you can tailor the dashboard to that audience's requirements. One dashboard with information that's relevant to some users but not all isn't as useful as two dashboards that only contain pertinent information for each audience.
- Omit irrelevant information: Only include information that users need to know. Information that's relevant but not essential is just distracting and decreases the dashboard's usability.
- Keep it simple: Resist the urge to add unnecessary visual elements and media to your dashboard. Keep the number of visual elements in a dashboard to less than 8. The main objective of a dashboard is to present information clearly, and unnecessary visualizations just clutter dashboards. Keeping dashboards simple will also ensure smooth performance.
- Provide context: Stand-alone numbers are often not that useful for understanding your operational status. Contextual information, such as target or historical values, make current values more meaningful and help users decide whether a situation warrants attention. Context can also assist users in identifying concerning trends, which can lead to actions that prevent problems (Few, 2017). You can provide context in your dashboard by setting reference values in indicators, including guides in serial charts, or using charts that are more appropriate for showing time-series data, such as line or area charts.
- Choose the right chart: Various chart types and styles are supported. Using the appropriate chart for your data will help make what you want to show clear to users.
- Notify users of important events: Configure the dashboard so that users are immediately notified when something occurs that requires action to increase your organization's responsiveness. Ideally, include no more than two notifications in a dashboard, to avoid flooding users with non-critical notifications (Few, 2017). You can draw attention to updates in a dashboard by using conditional formatting. This causes an element's appearance to change based on your data. The indicator is one of the elements that supports conditional formatting. For example, you can configure it to turn yellow and include an exclamation point when the number of crimes has increased by more than 20 compared to yesterday, and appear with a white background the remaining time, so that you can easily tell whether the number of crimes has significantly increased or not.
- Arrange and size elements appropriately: Organize and size elements based on their importance and relationships with each other. The elements with the most important information should be the largest and have the most prominent locations in the dashboard (Few, 2017). Place the most essential elements in the top left corner of your dashboard, and the least essential in the bottom right, since most people read from top left to bottom right. Also, group highly related elements so that they always appear next to each other and it's easier to see their close relationship. The map and map legend elements, as well as the list and details elements, are commonly grouped together.
- Use color wisely: A common design problem with dashboards is that they have too much color. This can make them visually overwhelming and difficult to read. Only use color to display data differences, and when other methods to do so aren't as effective. When you do use color, take advantage of the default color sets available in certain elements, such as serial and pie charts. The colors in these sets are designed to look good with each other. Another common issue is to use red to symbolize bad statuses and green to symbolize good statuses. This is a problem because for 10% of males and 1% of females, red and green look the same, due to color blindness (Few, 2017).
- Consider where the dashboard will be used: Think about the environment in which the dashboard will be used when designing it. If the dashboard will be used in a dimly lit office, consider using the dark theme to make it pop more. If the dashboard is going to be displayed on a monitor wall, as opposed to a desktop machine, it should be easy to read from a distance. Last, you should author a dashboard to look good at the screen resolution and zoom levels at which it will be displayed. In general, it's recommended to test your dashboard in the workspace it will be used in before sharing it with users.
To get started authoring your own dashboard, see Create a dashboard.
Few, S. (2007). Dashboard Design for Real-Time Situation Awareness. Retrieved August 1, 2017, from http://www.perceptualedge.com/articles/Whitepapers/Dashboard_Design.pdf