Oracle Business Intelligence and the Oracle Database Power Better Analytics
By Forumsuser10819-Oracle on Sep 23, 2009
By Plinio Monteiro, Director, Customer Support, Analytics
Analytics is supposed to answer questions that will lead to actions. Oracle Business Intelligence Enterprise Edition Plus is the most powerful tool available to help customers answer critical business questions, such as:
- How many orders were placed
- What customers are responsible for the highest volume of orders
- What they're implementing
- What exactly is the issue with any products/services
- What are the unusual pattern changes that will lead to actionable data
- Sales trends by region
I could go on and on about Oracle Business Intelligence Enterprise Edition Plus's innovative capabilities and easy-to-use features. However, what I want to highlight is another Oracle product that is a tremendous complement to Oracle Business Intelligence Enterprise Edition Plus. I am talking about Oracle Database 10g and Oracle Database 11g.
Oracle Database 10g and Oracle Database 11g come with a wealth of statistical functions, limited only by the user's imagination in terms of what one can accomplish in the analytical arena. The analytical capabilities that comes with the Oracle database help businesses answer questions such as:
- What customers are most likely to default on the credit card payments (prediction)
- What attributes contribute the most to customer satisfaction (correlation)
- How many Service Requests should be expected for the upcoming months (prediction)
We decided to explore the analytical capability of the database to complement Oracle Business Intelligence Enterprise Edition Plus to predict and to establish cause-effect. One example where we utilized the power of both technologies was to predict the number of orders for a given company. So, the challenge was to predict the number of orders for the next three months so that the company could plan for production, staffing, marketing, stocking, and more. As you know, any prediction requires availability of historical data. The more historical data one has, the more reliable the prediction. The idea was to generate a new set of three-month predictions of orders on a monthly basis. The least squared model (prediction equation) to predict was used. This model relies on the intercept and slope of a regression analysis to predict. In this case, the model is a simple linear regression that can be plotted by using the database functions REGR_SLOPE and REGR_INTERCEPT.
We wrapped these two functions with a PL/SQL package for simplicity (we could choose to call the database functions directly from Oracle Business Intelligence Enterprise Edition Plus). Basically this package runs at zero hour of the first of each month, generating an output fact table with the predicted values. This table becomes part of the Oracle Business Intelligence Enterprise Edition Plus's physical, business and presentation layer. That way, Oracle Business Intelligence Enterprise Edition Plus can present this data by region, country, state and product. This provides a comprehensive forecast by different components of the business.
Also, because these algorithms reside in the database, response time is significantly better than trying to crunch a complex statistical model at the application level with frequent interaction between the application and the database. Oracle's decision to put these statistical functions at the database level was brilliant, to say the least.
What was done above is just a small sample of how one can add more analytical capabilities to Oracle Business Intelligence Enterprise Edition Plus. This combination (Oracle Business Intelligence Enterprise Edition Plus and Oracle Database) is a key component of any organization that wants to achieve a competitive advantage at a very low cost of ownership. Companies that rely on analytics for their business decisions can gain a competitive edge. For example, how to price a product to maximize revenue, how many seats can an airline overbook (Yield Management), how to price hotel rooms and ensure high occupancy with maximum revenue (revenue management) are all areas where analytics play a huge role for achieving competitive advantage.
This was an innovative way to utilize the power of Oracle Business Intelligence Enterprise Edition Plus and the Oracle Database to predict orders so that the company could make intelligent decisions pertaining to the ecosystem associated to orders.