There are hundred things going
over my head when I think of Data Warehousing and Business Intelligence (DW/BI).
But where to start and how to put it all down, I am still not sure.
Anyway, everything starts and
ends with data. Data is growing rapidly every minute. How can we convert this
data into useful information and use this data to add value to any business?
This is the basic purpose of any kind of data analysis. Role of DW/BI is to extract
the historical data, compare it with current data and perform any kind of analysis
that will answer critical questions for various businesses. I am going to try and put some context here. If
I am analyzing the accounting data of a firm, it will give me insights into
things like,
- Is there a way to reduce the expenses of the firm?
- If a firm has given away discounts for a week, has it really made a profit for the firm?
- If a subscription has been given away for customers, do they really stay after the subscription has expired?
And these are the questions I am trying to answer for my project as well. So I am pretty much surrounded by data and analytics for a while. These kind of questions can be
easily answered using data analytics, DW/BI.
Now that I have explained what is
“my understanding” of DW/BI, let us get a bit technical. We need the data in a form that is easy to understand
and provides fast query performance. This is achieved using a technique called
as “Dimensional Modelling”. It is quite different from Relational Database
Management Systems in the sense that, RDBMS looks to remove redundancy by
normalizing the data whereas dimensional modelling looks at simpler and faster
retrieval of data.
One of the ways Dimensional Models are implemented is using Star Schema. Quite honestly, I used to think star schema is the most
complex thing in the world. But once I understood what it was, it changed my
perception of BI to being slightly simpler than I thought. Though it is way too
early to pass this judgment. Star schema is all about facts and dimensions.
I am going to keep the discussion
of grains, facts and dimensions for the next time.
References:
The Data Warehouse Toolkit - Third Edition, Ralph Kimball and Margy Ross
http://www.greenbookblog.org/2012/03/21/big-data-opportunity-or-threat-for-market-research/
