Using BI to Create Value
Interestingly, in the past when one talks about Business Intelligence (BI), it usually revolves around building OLAP cubes to support business needs. The reason being the OLAP cubes presents summarized aggregate information, which analysts can then use to analyze trends in the business domain they are in. For example, they could easily make use of the front-end tools to constructs reports that presents revenue per quarter, broken by quarters. Or they could compare the revenue per quarter across different years.
Data mining for a long time had remained hard to use, due to the lack of good tools for building the data mining models as well as visualizing. Though there are many tools for doing a particular aspects of data mining, none could allow the models built to be queried easily. Neither could many of these tools allow developers to build applications (web or windows) around it easily. Most require you to buy their proprietary BI portal software etc. This inhibits developers flexibility in using the models in their day-to-day application.
Since the launch of SQL Server 2005, I think some of this is about to change. I feel it is about empowering users/developers - With the ability to make use data mining models in the applications they use/built. The learning curve for an otherwise complex task for making predictions have been reduced. Right now, you could easily build a proof-of-concept (POC) to share with IT management how easy it is to introduce some form of predictive ability or "intelligence" into the application. With this new found "power", hopefully analysts can make more informed decisions.
It is really about creating value from the gold mines (i.e. the data warehouse, the databaseas) in an organization. Hopefuly, from these gold mines, you can find trends, patterns, rules that would have otherwise been impossible to find using just OLAP alone!
Data mining for a long time had remained hard to use, due to the lack of good tools for building the data mining models as well as visualizing. Though there are many tools for doing a particular aspects of data mining, none could allow the models built to be queried easily. Neither could many of these tools allow developers to build applications (web or windows) around it easily. Most require you to buy their proprietary BI portal software etc. This inhibits developers flexibility in using the models in their day-to-day application.
Since the launch of SQL Server 2005, I think some of this is about to change. I feel it is about empowering users/developers - With the ability to make use data mining models in the applications they use/built. The learning curve for an otherwise complex task for making predictions have been reduced. Right now, you could easily build a proof-of-concept (POC) to share with IT management how easy it is to introduce some form of predictive ability or "intelligence" into the application. With this new found "power", hopefully analysts can make more informed decisions.
It is really about creating value from the gold mines (i.e. the data warehouse, the databaseas) in an organization. Hopefuly, from these gold mines, you can find trends, patterns, rules that would have otherwise been impossible to find using just OLAP alone!
posted by SQL Server Gems at
4:50 PM
0 Comments:
Post a Comment
<< Home