Data mart support
This month, we’ve been putting support procedures in place for the HR strategic reporting data mart, and addressing data questions from its first users.
We’re making sure that all queries are tracked and raised by calls in Unidesk, the University’s incident tracking system. This makes it easier to route calls to the right people to deal with them. For example, requests to add a new user or to fix a problem with the PowerBI platform are handled by the Service Management team, while data errors are forwarded to the data warehouse team.
Several of those data issues have been a bit tricky to pin down. This is to be expected because we found and fixed the simpler issues during project testing. It’s the difficult issues that slip through the net.
- Sometimes the issue arises because we made a mistake in the data warehouse. One set of symptoms arose because the warehouse wasn’t adding a record if it already existed, even though it had a different insert date.
- Sometimes the issue is in the source data, such as when the new HR system added a field that wasn’t in the old one, leaving lots of records with blank values. This seemed erroneous to the analyst but was actually intentional.
- Sometimes it’s a mismatch in assumptions or processes. For example, the design of the warehouse assumed that all changes in the original system were delivered as change items, but sometimes assignments are deleted completely. This meant we had to implement our own change capture procedure for this case.
- Sometimes the issue reported is a misunderstanding by the MI analyst. In one case, the report needed an extra filter on the employee type field.
We’ve been grateful for co-operation from the HR team and our colleagues who use the data mart in helping us to resolve all these issues.