Context is everything. To avoid confusion and foster good decision making, dashboards need to be designed for the context they are intended for.
Note: This guidance is a work-in-progress. To see our roadmap, make feature requests, or contribute, please go to carbon-charts GitHub repository.
Dashboards for presentation purposes show viewers the current status of KPIs relevant to the business problem. A good presentation dashboard provides a “big picture” view of the data while serving as a guide for the audience to decide what areas they would like to focus on and explore.
Priorize the importance of data, then assign clear visual hierarcy to them. The most important data should have the highest contrast and occupy the largest area.
Most people in the west read left to right, then top to bottom. This is known as the F-shaped pattern. Place elements that are most important at the top of the page, then follow the F-pattern for elements with decreasing importance.
Make decisions on what the user would be most interested in seeing. Non-critical information is provided on an “as needed” basis. Design to reduce cognitive load by stripping away anything that potentially distracts them from that goal.
Consistency is a good practice for all dashboards. Always use the same color for the same data set on all charts in the same dashboard.
White space sets elements apart or brings them together to distinguish its priority. It acts as a visual separator by guiding the users’ eye through the design, making sure to amplify only what’s important. It provides relief, breathing room, and an easier way to scan a page. According to research conducted by Human Factors International, whitespace increases comprehension by almost 20%.
Exploration dashboards are designed to help users interact with the data to discover insights and identify patterns. Examples of actions a user may perform on data include search, sort and filter data, roll up and drill down. Exploration dashboards are intended for people who have the time and inclination to look beyond a primary view and have to be very interactive.
Consistency in an exploration dashboard is mission-critical for effective and accurate data analytics work. Design for consistency in layout, color, labeling, units, and interaction. Below are some examples.
Link charts’ modifications, such as filter and zoom, so when user manipulates one chart, other charts showing related data sets are automatically updated to provide multi-dimensional views of the user’s action.
Use annotations to highlight trend, average, peaks and valleys, etc, to provide additional information that helps people interpret fluctuations in the data. Annotations should not obstruct the view of data.