What's the difference between a cube and a report?

Report - (Title II)

A report is typically designed to answer one or more specific questions.  For instance, 
  • How many people graduated from program X in 2010-2011?
  • Of those enrolled in program X in 2010-2011, what percentage were women?
  • How many new teachers from program X complete a mathematics endorsement in 2010-2011?
The best reports are easily consumable for a specific user to make a particular decision.  For instance, the Certs Issued Report was really designed to tell the Director of Certs whether s/he needed to ask for additional resources (the number of actions went up or down).  The challenge of the Cert Issued Report was when folks attempted to appropriate it for their own needs, such as looking at the number of people completing from program X.  The Certs Issued Report had some useful insight to completion, but the numbers were a bit off.

Cube - (Completer Table)

A cube (OLAP Cube) can be thought of as is an intermediate step between the database and a report.  A cube works best when you have an idea of what you need to know for a decision, but you're not exactly sure.  Such as;
  • Are their differences in the WEST E tests pass-rates by a demographic characteristic?
    • Are these differences getting worse over time?
    • Are these differences worse from one institution to another?
    • Did this intervention help?
Databases will carry this information in multiple places (person information, test information, intervention information) and a cube will be a usable extract from the larger database.  Using a cube is not difficult, but it does take some level of training and understanding.  Interpreting a cube requires knowing the rules of how the cube was constructed.  For instance, on the completer table it is imperative that you know that it is counting completion actions taken (often endorsements), not people.


Pros and Cons

A single report is usually cheaper to produce than a single cube, especially when you consider that the cube will need additional analysis to find the actionable data.  Also, a report is almost always rolled up information, which means that there are seldom privacy concerns to address.  A report will also be easier to produce across institutions because there is less need to get all the definitions and manuals in place (but this also means that they can be inaccurate).

A cube contains data for many reports.  For instance, the completer table will answer many questions related to demographics and degrees, but it can also be linked to other data sources to answer even more questions to what has happened in the teaching workforce over time.  Creating the cube from a database is cheap, it's the resources related to the skills to consume that can be expensive.

It is possible (but not efficient) to create a report by hand.  It is much more difficult to create and manage a cube by hand.


Notes

  • Cubes are not databases.  In fact, they often end up breaking the rules of a database. 
  • Cubes are designed to make multiple reports, many of which are unknown prior to creating a cube
  • Even today, most of the data we receive from OSPI comes in the form of a cube (is not usable right out of the box).