/* ---- Google Analytics Code Below */

Saturday, December 27, 2008

Defining Business Intelligence Terms

In the BeyeNetwork Colin White looks at the various implications of the business intelligence terms: Data, Analysis and Visualization. He covers a number of variant views depending on the perspective of the writer. He uses a recent review of a recently introduced BI tool. Thought-provoking:

Business intelligence terminology is often confusing, and different people use terms in different ways to match their perspective on the market – and sometimes even to suit their marketing messages. Colin White discusses with industry specialists their perspective on data analysis and data visualization, and the need for the industry to have a common understanding and definition for these terms.

I like to see these as two key elements and not three. First there is the architecture of the data, and all of the meta data needed to understand it in context, and second are the models we use to find patterns to simplify data.

Models can be of many forms but generally are of two types: mathematical, which typically include logical or statistical methods. Second are models of design where we cannot directly quantify a result, but use our own visual systems to simplify a system. In both kinds of models they must simplify our understanding of a system, otherwise they have no value. They may or may not be predictive depending upon the relationship between the data, its context and the models used.

And always, the simpler the model the better, something that can be put on the back of the envelope is best of all.

No comments: