So you'd have point in time and all the other things you might get out of a data warehouse. I can see where this relates to using the dump(s) for an AdventureWorks replacement, with the additional dimensional or 3NF model (would be good to show the different modeling philosophies as well) updated to show common data warehousing techniques.

Because the Stack data is not particularly heavy on numerical/financial data, I'm not sure how good of an example it would be, but we do see regular questions on dimensional models on SO, so I thought I'd ask here.

As an (unrealistic) example, I currently use a data warehouse for customer profitability analysis. We have income on a customer basis and then allocate expenses through business units. In the Stack case, one might allocate the site revenues and expenses to customers in some way and determine what kinds of customers are most profitable for the site. This is done on a snapshot monthly - this doesn't necessarily require any different data model, but to run it over time, you do require monthly slices. Obviously, for something like that, you'd need click-stream data which is a LOT bigger than the SO data, and this would probably need to be modeled for efficiency.


No plans at the moment, I think we should focus on the SQL part of things for the near term, maybe down the line when SEDE is highly polished we can look at something along these lines.

| improve this answer | |

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .