Where Do OLAP, Data Mining Fit?

Online analytical processing (OLAP) and data mining are among several technologies that businesses use to gather, store, query, and analyze historical data to support business decision-making. OLAP applications are typically used to provide answers to questions related to historical business performance. For example, a grocery chain manager might use an OLAP application to chart sales for various product lines, locations, and time periods to answer the question, “What are the unit sales and revenues of each brand of paper towels sold in our stores during 2006?”

Data mining, on the other hand, uses pattern recognition algorithms to divine patterns or trends in data. These patterns and trends are used to build predictive models for specific business processes, such as forecasting sales, selecting a subset of customers who are most likely to respond to a direct mail campaign, or determining which products are likely to be sold together.

OLAP and data-mining systems typically work with data in data warehouses, large relational databases that serve as repositories of summarized business data collected from various operational databases throughout the enterprise. Data warehouses are commonly populated by extract, transform, and load (ETL) tools, which pull data from online transaction processing (OLTP) databases—such as databases that record individual retail sales transactions—and cleanse and summarize those data before loading them into data warehouses.