on - edited on by
DRay
This page is intended to keep a focus on Dashboard design in terms of:
- Planning
- Implementation
- User experience and functionality
- Dashboard training for end users
Dashboard Planning
Dashboard planning goes hand in hand with project planning, and will ensure an efficient, optimal and smooth project implementation.
Dashboard Implementation
The first stages of Dashboard implementation should initiate by testing and verifying data in early EC development stages. These tests should give initial immediate indications that the data is as expected, whether on complete data sets, or test tests, and should be done with straightforward widgets:
- Numeric Indicators - for complete totals, averages and counts
- Pivot tables - for segmentation and dimension table connectivity
- Pie charts - for ratios
- Trend charts - for time trends
These initial widgets can be an excellent base for initial OBP demos, for teaching basic to advanced front end features such as filtering (basic and advanced), measured values, data security multi-pass aggregations and much more.
After completing and verifying initial tests the initial test widgets can be a basis for creating formulas, and for the initial dashboard altogether.
Dashboard view
Two main dashboard design thumb-rules:
- Top to bottom - Left to right
- Cognitive research proves that upon initial view of written and visual material, the eyes will tend to focus at the top left hand side and move to the bottom right, as when reading a book, the most important KPI's should be shown at the top left hand side, and deteriorate with meaning/importance as you move to the bottom right.
- Publish and release in iterations
- Don't hesitate to share and publish dashboards right at the beginning of your development. You will always be able to introduce additional future requirements and releases by re-publishing dashboards, this will be transparent and seamless to the end users. This will also enable you to receive feedback and iteratively improve and optimize your dashboards.
- Certain KPI's together with their respective visualizations should be formatted with the end user's point of view and use case. The use case too is to be planned and conveyed to the user during dashboard usage training. A dashboard should have a structured view with immediate values at the top, moving down introduce certain trends which too can be immediate indications for monitoring, and higher resolution informative detail visualizations and the bottom. Remember not to overload on the end user with too many visualizations, and introduce user experience plugins when needed.
Choose your visualization, Data visualization Wizard
User Experience and functionality
Throughout the entire BI project implementation, during both EC and dashboard, one must think of the end user's use case and user experience. (For example, the manner in which a user will filter will determine the difference between standard text columns, to entire dimension tables in the EC)
Following Sisense features, and more, should be discussed and considered during first stages of planning and development, in order to see the end user's ultimate experience.
- Hierarchies - guide your user through their drilling and investigation, provide them with a guided path of drilling into their data and consider disabling 'drill to anywhere', preventing them from drilling to illogical paths
- Data security - Understand your disparate users' data's sources and structure. Disparate data can be consolidated and distributed among your various users and user groups using data security.
- User groups - Apply data security rules, share dashboards and apply additional processes on large amounts of users by use of group
- Star formulas and filters - Create predetermined formulas and complex features for future use of your dashboard designers.
Dashboard functionality plugins
Various plugins provide much enhanced functionality for Sisense web highly benefiting the overall user experience:
These plugins improve UX for a few reasons:
- Show more with less! - Instead of displaying one extremely robust and overloaded dashboard with many many widgets, give the user a birds-eye view of the initial information and enable drilling and previewing of much more data
- Take optimal advantage of valuable dashboard real estate
- Load upon demand! - with an accordion plugin, eventually, you'll be providing many widgets from various dashboards on one, BUT not all widgets will load upon first refresh/load. Only when the accordion/jump to dashboards are called upon, will the widgets require resources to load.
