What is a Dashboard?
We begin with a definition.
Unlike a single visualization, a dashboard presents multiple visualizations of related data together, often with interactions among them. Related measures, categories, and time periods can be viewed together without needing to switch among separate reports.
The purpose of displaying multiple visualizations simultaneously is that they all contribute to a common analytical theme. For example, consider sales analysis. We may want to view sales by geographical region, by product, and over time. These visualizations can be placed on a dashboard along with summary values such as total sales, average sales, and profitability.
Dashboards have the obvious advantage of displaying multiple visualizations with commentary in the same window. They also can be designed so that different visualizations interact with each other and with the user. For example, consider the following dashboard constructed with R that simultaneously shows multiple relationships and how those relationships change over time.
Dashboard: life expectancy, GDP, population, and geographical region
The static image appears in Figure 1, but traveling to the above link leads to the fully interactive version of the dashboard.
Drag the slider in the middle of the dashboard to the right to move forward in time from 1952 through 2007. The scatterplot then displays the relationship between GDP per capita and life expectancy across four major geographical regions as it changes over time. The dashboard supports two comparisons at once: changes in the relationship within each region over time and differences in the relationship across regions.
As the slider moves to the right, both GDP per capita and life expectancy generally increase for all four regions. However, the extent of these increases differs by region. The dashboard brings these related views together in one display, making comparisons easier than if the same information were spread across multiple separate screens.
Also notice that the top of the dashboard, in addition to the title, contains two tabs. One tab, Indicators, is for the visual display. The other tab, Data, allows the user to directly view the data from which the visualizations are constructed. The dashboard itself can consist of multiple pages or tabs that can be viewed independently.
Designing a Dashboard
To construct a dashboard, begin with a specific goal for the information to communicate.
Each visualization should contribute to the viewer’s understanding of the same question or related set of questions. A dashboard is not simply a collection of charts. The visualizations should work together to support comparison, explanation, monitoring, or exploration.
Related views should be placed near each other, use consistent scales and colors, and support visual comparison. Interactivity should clarify the analysis, not complicate it. Filters, sliders, and selection controls are most useful when they help the viewer move from overview to detail, as with the Time slider in the Figure 1 dashboard. As with any visualization, a dashboard should avoid clutter, emphasize the most important information, and provide enough context for the viewer to interpret the display.