Well, what is a dashboard? At first, it simply looks like a set of reports nailed together on a page However, this misses an important point about dashboards, which is that they give the viewer something which is ‘over and above’ the individual reports give to the data consumer. A dashboard can mean different things to different people. There are a number of different types of dashboard, which are listed here:
Strategic Dashboard – overview of how well the business is performing
Faceted Analytical Display – multi-chart analytical display (Stephen Few’s terminology) This will be discussed in more depth next.
Monitoring Dashboard – this displays reactionary information for review only; this data is often short-term, perhaps a day old or less.
Each dashboard type has got the following elements in common:
- Dashboards are intended to provide ‘actionability’ in addition to insight; to help the data consumer, to have insight into the presented data.
- The reports on the page support a ‘theme’, which is the fundamental business question which is answered by the dashboard. In other words, what is it that the business need to know, and what is it that they need to act upon?
- Further, the dashboard should rest on a fundamental data model, which has data that is common to all of the reports; the reports should not be completely disparate. If this occurs, then the data’s message may become diluted as distractions are added.
In order to explore the idea of the Faceted Analytical Display, I have used data from Microsoft Connect items, which are focused on SQL Server Denali. This dashboard shows us different perspectives on the numbers, types and statuses of Connect items opened for SQL Server Denali. In order to understand more, it is possible to select relevant years on the right hand side, in order to show how the data has changed over time. If you click on the image below, it will take you to the Tableau Public website so that you can have a play for yourself!
Thus, this dashboard type is, in Stephen Few’s terminology, a “faceted analytical display”. Few defines this as a set of interactive charts (primarily graphs and tables) that simultaneously reside on a single screen, each of which presents a somewhat different view of a common dataset, and is used to analyse that information. I recommend that you head over to his site in order to read more about the definitional issues around dashboard, along with practical advice regarding their construction.
This dashboard isn’t a straightforward ‘Monitoring’ dashboard, because it does allow some analysis. It is also possible to ‘brush’ the data, which means that it is possible to highlight some bars and dashboard elements at the expense of other elements. There are other considerations in the creation of the dashboard:
Colour – I used a colour-blind palette, so there are no reds or greens. Orange and blue are ‘safe’ perceptually distinct colours. At the foot of the dashboard, the same colours were assigned to Connect call status. So, ‘Fixed’ has the same colour for both ‘Closed’ and ‘Resolved’ connect calls, as this is the same for the other status types.
Bar charts – for representing quantity and for the purposes of reading left-to-right, and for facilitating comparison within dashboard elements.
Continuous data – the number of Connect items opened at any point is given as a continuous line chart. This line chart is interesting, since it shows that the number of Connect items has increased dramatically since the start of 2011. It’s great that everyone is getting involved by raising Connect items!
I will be interested in your feedback; please leave a comment below!