Data Truth or Dare: “Cubes”

November 07, 2013

Data Truth or Dare: “Cubes” – Are they a type of OLAP database OR a brand new kind of server rack?

Welcome to another exciting game of Data Truth or Dare. You can catch up on past entries here. This week’s question is about Cubes. Are they a type of database OR a brand new kind of server rack?

Data Truth: Cubes are a type of OLAP database that facilitates your analytics and are ready for slicing and dicing.

OLAP (Online Analytical Processing) cubes are a computer-based technique used to analyze data in search for business intelligence. The most common form of OLAP databases is the Multidimensional OLAP (MOLAP), which manages data in a cube structure. Each cell of the cube holds a number that represents some measure of the business such as sales, profits, expenses, budget, location, and year.

Picture a Rubik’s Cube. Each of those individual squares is a dimension stacked on top of the others to symbolize the data added up by every possible combination of dimension and hierarchy level. For example using the dimensions I mentioned above: sales, profits, and expenses added up by location, by year, by location and year, and every other possible combination. Those calculations produce a data structure that looks, quite literally, like a cube.

The numeric values in your data (such as revenue and profit) are placed in the intersections in the cube structure. The value in the cell is the aggregated value of the dimensions (such as customers, period, and product) that meet in the intersection. As the values based on different dimensions and aggregation level is pre-calculated, the use of OLAP cubes significantly reduces the query time compared to relational databases.

Interested in learning more? Have a look at the free TARGIT ebook, the BI Dictionary, to hone your skills for the next Data Truth or Dare.