Sources of Data
In designing the way that data flows through your enterprise, one of the first steps is to consider where you are obtaining that data. Traditionally, businesses have only thought of one data sourcetheir on line transaction processing (OLTP) systems. This is still a significant source of data, but it's important to recognize the variety of other data sources from inside and outside the organization.
On Line Transaction Processing (OLTP) Systems
All the detailed transaction information that is generated in OLTP systems can be used to understand what is happening in a company.
Web Clickstream Data
Web clickstream data is the record of Web site activity. It includes information about customers, where they're coming from, what they're looking at, and where they're going.
The use of DTS in transforming clickstream data is discussed in Chapter 1, "A Quick Look at DTS."
Data from Outside the Organization
More enterprises are using data from outside the organization. Address lists can be purchased to help cleanse customer data. Census data, market research data, and potential customer lists can be purchased to assist in the process of making business decisions.
The Data Warehouse
A corporate data warehouse is a destination for data transformationa place where data is gathered from multiple OLTP systems and integrated into a single, unified perspective. A data warehouse is organized into functional areas, and it provides a historical perspective on corporate operations.
The data warehouse is also a data source. Because its data has already been unified, homogenized, and cleansed, it is often the best source of data for business analysis.
Results from Business Analysis Systems
Business analysis has always affected the ongoing operation of a business. Managers look at the data and make decisions about future operations on the basis of that data.
One of the current emphases in the field of data warehousing is closed-loop business analysis. In a closed-loop system, the OLTP data is brought into a data warehousing system, it is analyzed, and the results of the data are sent back to the OLTP system to automatically modify the transaction processes.
Here's an example of closed-loop business analysis. A data mining program analyzes which set of customers is likely to want to buy a newly published book. This information is automatically transferred to the OLTP system, which sends out information on the book to the selected customers.
Figure 3.6 shows the flow of data from the various sources to the analysis systems and then back to the OLTP system.