Databases Reference
In-Depth Information
same tasks continue to be done, even after the end of the project. Some of the
steps that are repeated include the following:
Building report templates:
This involves setting up the report layouts,
exception criteria, and parameters. It also includes developing how the
results are presented, such as a tabular report, a chart, or a scorecard.
Building the initial reports is done by a BI developer, either from IT or
the business.
Creating additional reports:
Once the BI application is released for use,
any other users who have been granted appropriate privileges can create
other reports.
Setting up navigation:
Building the navigation paths is the work to
organize and display how the starter set of reports can be located and
accessed. This work is also done by a BI developer, but it can be a bit
more technical. Often this is done by a BI developer from IT, but it could
still be completed by a BI developer from the business community. This
work is not typically done by other business users of the application.
The project work to build the BI application is almost done when the reports
have been built and the navigation paths are set up. Now it is time to test
everything, a very important step that must be done to ensure that the BI
application works correctly.
Testing BI Applications and Validating Data
The BI application must be tested in a similar manner to any other system
that is tested. How the data has been defined in the BI tool must be validated.
This includes ensuring that drill paths are set up properly and that business
measures are being calculated correctly. Then, each report must be tested
to ensure that it delivers what was intended. Some reports may be created
properly but may not provide the expected insight. It is worth adapting these
reports to be more useful, with the guidance of the business community.
The second type of testing that is done during BI application construction
is validation of the data itself. Regardless of the extent of the testing and
validation that was done during construction of the ETL system, more data
problems will be discovered now. Due to the visual nature of BI tools, data
problems are often quickly identified. A lot of this data validation occurs during
the development of reports. As problems crop up, work with the entire project
team to figure out the source of the problem and work toward a resolution.
Data problems may result from errors in defining business measures in the BI
tool, errors in the ETL processing rules, and sometimes even in the underlying
source system. The investigation of data problems can be complex and time
consuming.
