The Data Mining Dream
-
Business Intelligence is an exciting idea. I think most people are inherently attracted to the dream that we can make sense of the morass of data that drives our lives, to understands it and use it to "predict the future".
Data mining projects are springing up everywhere, digitizing and mining everything from the Library of Congress, to 80 million documents from the Guatamalan police archives, in an effort to uncover human rights atrocities. Alongside this, BI technology strides ever onwards. Heck, this week Business Objects even announced a new set of business intelligence tools aimed at "companies with less than $1 billion in revenue". Hands up whose firm falls into that category? At last, BI for the little guy!
.Seriously though, I understand why Sean McCown is excited that Donald Farmer is excited by SQL Server BI. He wants to see intelligent and predictive analysis tools made available to DBAs, so they receive warnings about query performance degrading, and so on, before it happens. At this level BI is exciting and tangible.
But at the broader level, of being able to mine for consistent and valuable information all of the data that comprises every arm of your business - to most companies this is still a pipe dream. So what is holding BI back? Technology? Cost? Fear of "transparency"? Those last two are probably a factor, but isn't the real issue the simple fact that it's impossible to know what you'll really get out of it - and unless you have, right from the beginning, a fundamental, shared understanding of the meaning of your business data, then it could actually be very little. In his recent data dialog, Phil Factor mourned the demise of the grizzled old data architect/systems analyst whose job it was to establish the shared knowledge of what all of the entities that comprised the business data actually meant. If you don't have that understanding, if your data is fundamentally unsound, then no amount of manipulation and mining is going to help you.
-
As a systems analyst/information analyst who was unaware that they had demised i think there is a fundamental barrier for many small to medium sized businesses when looking at BI projects, namely timescales. Our business model can shift in a matter of days to reflect changes in the market. It's clear from reviewing data from just 12 months back that predictive analysis would have been a waste of time and using it as part of the data set in an analytical process to inform business decisions now would give us completely spurious results. The upshot is that building a comprehensive tool would take at least a year with my available resources and would be completely out of date when complete.
K.
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply