How have your organizations handled the control and ownership of Derived or compound datasets created for analytic purposes? When Derived or compound datasets are constructed from data owned by several information owners, who or which group provided approval for access to the object? Who or which group managed changes in terms of requirements for the derived or compound object?
Sort By:
Oldest
Chief Data Officer in Software9 months ago
This is a really great question, and in my experience, most companies would not be able to answer it. In theory any new derived or inferred data that had not existed previously should be reviewed and approved by some governance committee, but in reality that rarely happens since often the governance committee wouldnt even know about it. Some form of data catalog that's actively profiling for new metadata could help here, but its a reactionary process. Chief Data Officer in Media8 months ago
Malcom already provided a great answer to this one. I would just like to concur that ideally, a data governance committee should oversee the creation, access and use of derived or compound datasets. A data catalog can help to create periodic audits and reviews of any dataset used by the organization or having a tool that offer a semantic layer on top of all environments, but in reality these datasets can also fly under the radar. So it's tricky. Having policies and a well established data governance program can help, but in reality it's challenging to manage derived/ compound datasets