Flexibility usually comes with complexity. To facilitate the design process, wizards walk the user through the steps required to configure an element. Once created, an element can be re-configured using property dialogs. Similar to PerformancePoint workspaces, Dundas uses projects to organize dashboard pages and elements.
Out of the box, PerformancePoint includes data providers for Analysis Services, SQL Server, Excel Services, Excel Workbooks, and SharePoint lists. Dundas Dashboard lets you connect to:
- SQL Server 2005 and 2008
- Microsoft Access
- Oracle 10g or 11g
- MySQL 5.x
- Microsoft Analysis Services 2005 and 2008
- Microsoft Visual FoxPro
- MS OLAP
- Other data sources (CSV, Excel, SharePoint lists, Google Analytics, etc.) More information on Dundas data connectors here
Implementing Performance Dashboards
Dundas Dashboard defines two types of dashboard: performance and analytical. A performance dashboard is closer to how you may envision a digital dashboard. It has a fixed layout and typically displays a set of key performance indicators and other elements. By contrast, an analytical dashboard lets users create interactive reports from an Analysis Services cube. Implementing a performance dashboard is a "click-intensive" process that involves the following main steps:
- Configure data connectivity – Set up one or more data connectors using the supported data providers.
- Create virtual views to abstract the relational or OLAP schemas - Similar to data source views (DSVs) in Microsoft BI projects, Dundas supports virtual tables and cubes that are especially useful if you need to combine data from multiple tables in a relational database. When targeting an Analysis Services as a data source, however, I've found this layer somewhat redundant and I hope a future release deemphasize it.
- Create dashboard elements – There are two types of elements for performance dashboards: datasets and KPIs, which I will explain in more details in the next section.
- Assemble the performance dashboard – This step involves adding dashboard elements, defining parameters, and adding interactive features.
Now that we covered the foundation, let's take a closer look at the building blocks of a Dundas dashboard starting with datasets.
Figure 2. Tabular data can be defined as a dataset.
As its name suggests, a dataset is an element that lets you visualize a tabular set of data, such as in a grid layout. During the process of defining a dataset you specify measures for the numeric data, dimensions that will be used to group the data, and sorting and filtering options. During the dataset configuration steps, you can assign a preferred data visualization, such as a chart or a data grid. Figure 2 shows a dataset visualized using the Data Grid control on a dashboard page.
Once the dashboard is rendered, the user can click the column captions to sort the dataset interactively. What’s more interesting is that you can specify a refresh interval to auto-update the dataset when displayed on the dashboard page.
Key Performance Indicators (KPIs)
Although Dundas Dashboard doesn't force you to define key performance indicators (KPIs), a dashboard would typically include one or more KPIs to let management track the organization’s performance. If you use Analysis Services, you can take advantage of its great support for KPIs and define them in the cube. PerformancePoint lets you import the KPI definitions from a cube and use MDX functions, such as KPIValue, KPIGoal, etc., to retrieve the KPI metrics.
Unfortunately, version 2.5 of Dundas Dashboard doesn't provide native support for Analysis Services KPIs and they are not directly accessible. Instead, you need to map the KPI properties to calculated members or measures which is a good idea anyway if you want to have more control over the expressions, such as to specify format settings. Once this is done, the KPI Setup wizard walks you to the steps to configure a KPI, including the measure that will be used to derive the KPI value and additional contextual measures for the goal, status, trend, etc. Interestingly, you can also specify dimensions for slicing the KPI if you prefer to visualize the KPI in a chart or set up a filter.