So we know we have data but what does it look like and how can we find out more about it? With extremely minimal configuration, the Metadata harvesting feature of Information Steward enables users to gather data about their data - quickly. When you run a Business Intelligence (BI) report, and the data looks wrong, the turnaround time for IT to analyze and potentially provide a workaround could mean a loss in revenue. And sometimes there may be no real data integrity issue but rather a lack of understanding on how the data is derived. What if we had an environment where IT and the business could collaborate and eliminate some of this lead time? Your business could spend more time generating revenue, and IT could spend more time on innovation.
Setup an Integrator Source
To get started, you need to set-up an integrator source to harvest metadata from your BI environment. In this example, we'll harvest metadata from a Business Objects reporting environment. The configuration for integrator sources is neatly embedded in the SAP BI Platform landscape, and it is easily accessible via the Central Management Console (CMC):
Integrator sources can be easily configured and saved in groups if required:
Verifying Data Lineage
After the integrator source has been run, SAP Information Steward is then able to provide critical lineage information that enables your business users to easily and quickly understand the source data for their reports. This gives IT a great start on both analyzing the problem and educating the business users about their data. In the example below, we see how dashboard data is derived - all the way back to the physical database table - by entering the dashboard name 'efashion' in the search field, selecting the dashboard and viewing 'Lineage'.
The very nature of a BI architecture means that the smallest of changes at the technical level can cause catastrophic problems for many business users. Think of it as the 'Butterfly Effect'! :) But how do you mitigate this? Do you plow through reams of ETL logic? Do you write code to automate the understanding of report sources? Truly mitigating this problem can be time consuming! Information Steward can help IT understand the cross system impact of a simple database field change. Similarly, by entering the database table name 'shop_facts' in the search field, we can assess impact and see what reports could be affected by a change at the database layer.
In addition, Information Steward enables you to link the metadata shown above with additional relevant metadata such as associated ETL processing. Here we can see a multitude of objects associated with an ETL Repository which can all be analyzed independently:
Looking Ahead to the Next Post
Additional features, such as 'Scorecarding', are seamlessly integrated into the Information Steward product. Tune in to the next post in the series to learn more about Scorecarding!