1
$\begingroup$

I am reading the book The Data Warehouse Lifecycle Toolkit by Ralph Kimball. I come across the term Stovepipes fairly often. After doing some research I read that Stovepipes are when you don't have conformed dimensions to link data marts. This short description was all I could find every other resource I looked at simply said that Stovepipe models are bad.

What is is meant by "conformed dimensions"? What exactly are these Stovepipes and can somebody provide me with an example?

$\endgroup$
  • $\begingroup$ These are the author's nonstandard terms. Accordingly, A dimension is conformed when two copies of the dimensions are either exactly the same (including the values of the keys and all the attributes), or else one dimension is a perfect subset of the other. See also dinesql.blogspot.com/2016/07/… $\endgroup$ – Emre May 11 '17 at 18:09
  • $\begingroup$ Hi, if my answer clarified your understanding of these terms, may I ask you to mark my answer as the right answer by clicking checkmark at the top right of its body, just below the score? :) $\endgroup$ – VividD May 15 '17 at 16:57
2
$\begingroup$

It seems "stovepiping" is a term that legendary author on the subject of data warehousing Ralph Kimball adopted, and it looks that it actually originates from intelligence community. In its original (intelligence community) meaning, it is mostly related to the usage of raw data, instead of processed data.

From Wikipedia:

Stovepiping (also stove piping) is a metaphorical term which recalls a stovepipe's function as an isolated vertical conduit, and has been used, in the context of intelligence, to describe several ways in which raw intelligence information may be presented without proper context. It is a system created to solve a specific problem. The lack of context may be due to the specialized nature, or security requirements, of a particular intelligence collection technology. It also has limited focus and data within is not easily shared.

From RationalWiki:

Stovepiping is a term originating in the intelligence community to describe a process by which raw data is funneled directly to high-ranking officials or the media.

Also, there is an interesting article in PhraseFinder.


Even better explanation from Relational Solutions Blog:

Most data marts are designed as one off reporting solutions, designed to solve an immediate problem that the business users need to solve. When designed as a stand-alone, they are often referred to as “stove pipes” or “silos” of information.

Analysts new to the space think these are new terms, but Data warehousing consultants have used these terms since the 90’s. They’re used to describe stand-alone reporting solutions. Typically these stand-alone solutions are developed by individual teams or departments.

These groups develop “silo’s” or “stove pipe” reporting databases to achieve a specific goal that they were unable to get financial approval for. If they have a need for something that you can't get approval for, you resort to building something on your own. It happens in every company and every department.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.