I came across Vincent Rainardi’s blog, and thought I’d share it with the readers.
I’d also like to thank Vincent for writing up the review so quickly.
I browsed through Vincent’s blog, and found that he authored this book:
Many reviews praised that it is an excellent book on data warehousing that combines simple and practical explanations and "how to" for all the key aspects of building a data warehouse on SQL Server.
I came back to my old job, but am not working on exactly same projects. I am very happy that I have new puzzles to solve now.
Obviously employee compensation and productivity reporting has become a hot topic in the business recently.
I am going to blog on this subject as I get to understand more about the business needs.
After completing several ETL processes and reports so far, I gradually start to put all the puzzles together.
In its simplistic form, I can envision a subset of the data warehouse bus matrix. This matrix will revolve as I understand more.
The dimensions will involve hierarchies. All the facts basically are from a transactional table. But as we put together all the fact tables for reporting, we will get all three types of fact tables, i.e. the transactional fact tables, snapshot fact tables, and also accumulating fact tables. The third type will be the most common.
The natural granularity for our business transactional fact table will be one row for each action a team member does for each account. For the purpose of employee compensation and productivity reporting, the facts associated with the business process typically just include the counts of the actions, such as the counts all “declines”, or the counts of all “approvals”.
In the next blog I’ll talk about whether I need to normalize my fact tables by fact types.