The process for developing a dashboard is the same as developing any other business intelligence tool. At a summary level, this includes:
- Identify key business strategies
- Link performance KPIs to objectives supporting the strategy
- Identify the appropriate location to measure KPI’s
- Implement measures and collection of data in appropriate time intervals
- Store data in a central data repository
- Design and develop dashboards for each required role and business area
- Integrate use of the dashboard into standard operating procedures **CRITICAL**
- Distribute dashboard to users, using login business rules to define view and data portlets
- Implement dashboard QA
- Monitor and measure dashboard performance
Effective dashboard design:
Trend #1: More analytical dimensions for executive dashboards
One obvious area for improvement in executive dashboards is in the number of dimensions (in the analytical sense) that can be represented on a two-dimensional surface (in the physical sense) such as a computer screen. In a typical bar chart, we see a numerical measure (say, profit) sliced by the members of a dimension (for example, the month members of the time dimension).
Trend #2: The demise of 3D for enterprise dashboards?
The first is simply occlusion (the data at the front has a tendency to obscure the data at the back), and the second is providing humans with controls that allow them to easily manipulate three-dimensional space.
Trend #3: More interaction for enterprise dashboards
Another major differentiator between paper and computer is that the latter can provide a high level of rapid interaction with the data. This area seems primed to offer the greatest source of innovation in the future. The integration of data visualization to allow regular business users to explore and ask questions of large amounts of changing and diverse data represents the future of BI. In other words, dashboards should simply be a starting point from which users can dig deeper into the data, and the display of the information should be so fast that the user is never aware of a delay. The trick here is twofold: The database engine behind the dashboard has to be lightning fast, providing the data as soon as it is requested, and the user interface has to be highly intuitive.
Trend #4: Focus on executive dashboard design and usability
To get to better usability, we need to get better at design. Indeed, however hard it may be for some technical people to admit, the main problems we face in turning data into information are not technical. We know how to handle data as dimensions and measures; we can produce very fast databases. What we also need is an interface that is so easy to use that the user is essentially unaware of it. There is no magic bullet here: We are still learning how to make interfaces to complex data easy to use, so this is an area where we should see significant advances in the future. The good news is that people are working on the problem and coming up with innovative ideas.
Trend #5: The importance of collaborative visualization in dashboards
And one final trend in dashboards: If we accept that it is a huge advantage to allow users to interact with the data (and it is), it follows that when they find something interesting, they will inevitably need to share it with others. And they won’t simply want to share a static view of the data, they will need to share the fully interactive view. This leads us to another area that Mackinlay believes is going to become crucial — “collaborative visualization.” Indeed, he argues that some analytical processes cannot be achieved by a single individual because they require the combined expertise of several specialists.
http://www.theiqx.com/dashboards/Developing-Dashboards.php
http://searchdatamanagement.techtarget.com/generic/0,295582,sid91_gci1362219,00.html
Reflections:
Dashboards must be made effectively to serve its purpose. Less decorations(non-pixel data) is important. This is so as it is the delivery of the message in its content that is important and not the display of it. An effective dashboard is so only when data are conveyed in the most accurate and fastest way.
BI reflections 