Designing Data Dashboards

Simplicity is the guiding principle for effective dashboard communication. By their very nature, dashboards need to communicate a business reality in the clearest possible way. Only the information that users require should be included. Typically, this data applies directly to the employee's job function. Dashboard data should also say exactly what needs to be said, using a vocabulary that the audience understands and a level of precision that is appropriate.

The Four Properties

There are four properties of a well-designed dashboard.

1. Expert organization. Careful attention must be paid to the layout and organization of every element. Other parts of this blog entry will provide details about dashboard organization

2. Condensed summaries and exceptions. Condensed summaries such as sums and averages should be used to explain large sets of data. Exceptions and abnormal conditions must be noted and identified.

3. Specific and customized. Dashboards must allow the audience to see the information that it most requires. Consequently, dashboards differ by audience.

4. Displays with small media. Small media can convey information efficiently.

The Information-Pixel Ratio

One way to achieve design simplicity is to aim for a low information-pixel ratio. This concept is related to Edward Tufte's data-ink ratio as described in the statistician's award-winning book, "The Visual Display of Quantitative Information". Dashboard designers should strive to use a small amount of pixels to display the whole dashboard. Of these pixels, a high percentage should be devoted to the non-redundant display of data.

De-emphasize, Regularize and Enhance

Another way to achieve simplicity and clarity in dashboard design is to increase the speed at which information is perceived. This is done by de-emphasizing and regularizing the non-data pixels that remain. By de-emphasizing these non-data pixels, the most important information will capture the user's attention – making it noticeable at a glance.

Common dashboard items that are non-data pixels include:

· Large logos, headers, and navigation areas

· Eye-catching graphics, logos, and background decoration

· Different colors without semantic reason

· Unnecessary frames and different-colored backgrounds

· Shadows and color gradients

· Distracting grid lines

· Area gauges

· 3D charts

After ensuring that non-data pixels are de-emphasized, dashboard designers must enhance the remaining data pixels. To do this, all unnecessary data pixels and redundant information should be eliminated. Eliminating unnecessary precision in numbers is another step that can be taken.

The Keys to a Great Information Experience

Dashboard designers must emphasize the most important information. The two best places for this information are the upper-left corner and the dashboard's center. For extra emphasis, the information may be enclosed in a border. Adding stronger colors or a special symbol to data points are other methods for highlighting information.

Applying these design principles doesn't guarantee that users will get the most benefit from a dashboard. A well-designed dashboard must be coupled with a great information experience. After all, it is the dashboard that acts as the final tool for data analysis.

There are four keys to providing a great information experience with dashboards.

1. Choose metrics that the user can act on. Interesting information isn't good enough. Designers should develop dashboards that provide actionable data.

2. Draw attention to most urgent information. Trends are important, but the current state-of-affairs is critical.

3. Progressively reveal data as the user expresses interest. Dashboard users may not have time for all of the information at once. Reveal information on a "need to know" basis.

4. Different views for different audiences or perspectives. Remember that just as a well-designed dashboard is specific to an audience, so is the information experience.

Turning raw data into information stands at the heart of business analysis and reporting. Data comparisons and graphical representations can also help in dashboard design.

Editor's Note: This is Part 2 of a two-part series. Click here for Part 1.

Resources:

http://www.perceptualedge.com/about.php

http://www.gartner.com/it/about_gartner.jsp