Hi All,
I’m deciding on the data architecture for a dataset that provides data about the Hardware that my agency’s IT department manages (computers, monitors, huge displays, etc.). I’d like your input about whether there’d likely be value in creating multiple stars since there are, arguably, multiple business processes involved including deploying/retrieving, attesting, supporting, managing inventory and retiring. A simpler alternative would be to conceive it as one business process – managing assets, with the activities above being just ‘type of activity’ attributes. That choice might miss opportunities.
One reason I could see value in splitting the data into stars is that I could then join the smaller, tightly focused stars with other related datasets. For example, the repair star could be joined with incident (repair) tickets, the deployment star could be joined with task tickets for delivering the assets.
Another is that the inventory management business process, for example, could probably benefit more from a snapshot fact table, but the latter would be irrelevant to the other business processes.
This choice would probably be easier if I knew the analytical questions that will be asked of this data, but I don’t, so I’m guessing so that I can have data at least close to ready in anticipation of management’s needs. My management appreciates when I lead them in discovering what’s possible to do with data. I know, for example, that we need to know where inventory is stored and whether we have inventory in the correct stockrooms spread around our territory, Joining inventory with repair & incident data could help judge that.
Does anyone have opinions, and ideally experience with similar data?
Regards,
Julie