Olap - a Better, Faster Tool for Data Analysis

OLAP- A Better, Faster Tool for

Data Analysis

Mike Sawyer

DBM405

Stephen Jones

August 8, 2005

Raw data collected and maintained in a database is meaningless, unless it is assembled and analyzed in some fashion. One technology that has been designed to do just that is called OLAP. OLAP stands for OnLine Analytical Processing, and is defined different ways by different sources.

Webopedia.com defines OLAP as "...a category of software tools that provides analysis of data stored in a database." This simple definition tells us that there are multiple components to OLAP, and that it is used specifically for analyzing data that is maintained in a database. So what is it?

SDG Computing has a better way of describing it; they liken OLAP to "...A million spreadsheets in a box". They state that OLAP is really a spreadsheet program that allows an individual to examine data from multiple perspectives. It allows someone to view the requested data, and ask for different views based on any questions that are derived from the current view, making it fast, and interactive. So how does it do it?

First, an OLAP server is set up between the client, and the DBMS (DataBase Management System). According to Wikipedia, OLAP begins by taking a snapshot of a relational database, and then restructuring it into a dimensional database. Queries are run against this snapshot. Data is pre-computed and placed on different levels. Because of this restructure, the queries can be run much faster than on a relational database, sometimes in "...about 0.1% of the time to run the same query on relational data." In any business environment, speed and efficiency are always an advantage, making any product or technology like this worth consideration.

The data is displayed in a format called an OLAP cube. This format allows the user to run a query, and examine multiple perspectives simultaneously. These different perspectives allow the user to produce reports that are more detailed. A cube is multi-dimensional. By definition, a cube has only three dimensions, but an OLAP cube can have as many as 128. A dimension is a category that the user chooses to see his or her data calculated in. SDG Computing gives a great example. For instance, the user could select these dimensions to gather data:

Product

Time

Store

Customer Age

Customer Income

Employee

If the user wanted to break down the information he or she received further, they could go as far as setting up different columns for each age group, and a different row for each store, and have a table that displayed the age group of the different customers who were shopping at the different stores. Each dimension has different levels, called hierarchies,