The SAS BI Dashboard indicator data object associates
a data source with an indicator. The core object is the indicator,
and a dashboard is just a collection of indicators. An indicator never
has more than one set of indicator data (and is rarely used without
indicator data). Access to four types of data sources is supplied
with the SAS BI Dashboard:
-
SQL queries, which can access relational
data
-
information maps, which can access
relational data and OLAP cubes
-
tables, which can access data in
a SAS data set that is registered in SAS metadata
-
stored processes, which can access
various types of data
Note: The SAS BI
Dashboard administrator can add access to other data sources.
Before
you create a dashboard, you must understand how to create indicator
data. Understanding the data flow in the SAS BI Dashboard is the key
to building enterprise dashboards that operate efficiently within
your organization’s business intelligence system.
Unlike
the flow of data in a report (which is usually relatively simple),
the flow of data in a dashboard can be very different. Consider the
dashboard in your car. Although you see a single representation of
the state of the car, the state is actually a collection of different
types of data received by the dashboard. The fuel gauge receives data
from the fuel tank, the speedometer receives data from the wheels,
the battery gauge receives data from the battery, and so on. Like
your car’s dashboard, a SAS BI Dashboard can have disparate
data sources.
Whereas
a report created with SAS Web Report Studio might fill several screens
with data from a single information map, a dashboard might render
data in a small display that is the result of SQL and JDBC queries
and information maps. A dashboard can also render the output of stored
processes that produce static images.
By using
information maps and SQL queries to retrieve data, you are unconstrained
with how the data is laid out at the data set level. For example,
you can use computed columns and grouping in the indicator data. After
you have the data configured, the lack of constraints helps you to
get initial dashboards set up quickly. But this same simplicity means
that data structure is not enforced.