The key to any business success is in improving the decision-making process with the introduction of business intelligence; the quick and easy access to information that should be found in data warehouses with the potential for generating multi-dimensional queries.
Common functions of business intelligence technologies are: reporting, online analytical processing (OLAP), data search, discovery of knowledge in data (data mining), business performance management, benchmarking, text mining and predictive analytics.
In this article, we’ll describe the technology of interactive analytical processing — OLAP, which is a relatively young technology with great potential for application in the business environment of start-ups and small to medium-sized enterprises.
What Does OLAP Mean?
The concept of OLAP is based on the principle of rich data representation. In 1993, the term OLAP (online analytical processing) was introduced by Edgar Codd. Taking into account the shortcomings of relational models, he indicated the impossibility of connecting, viewing and analyzing data with respect to the multiplication abilities (which are required in corporate analytics). Thus he defined the general requirements for the OLAP system — increasing the functionality of relational DBMS and introducing multi-dimensional conceptual view.
In a large number of publications, the abbreviation OLAP stands for a rich data view and data preservation to compose a rich database. Codd declared that “relational databases are the most suitable technology” for collecting corporate data. There is no need for new DB technologies, but rather, to complement the basic DBMS functions with data transfer and automated data analyses.
OLAP System — Functions & Features
Together with the multidimensional conceptual view — the OLAP technology possesses the following features and functions:
- Slice and dice technique — slicing takes a single value from one of its dimensions and creates a subset in the cube (3-dimensional array of data). For example with the sales cube, we can slice out certain years in a new data cube. Dicing choses certain values of the multi-cube and results in a new sub-cube.