DATA WAREHOUSING FUNDAMENTALS: A Comprehensive Guide for IT Professionals

15.2. DEMAND FOR ONLINE ANALYTICAL PROCESSING

Recall our discussions in Chapter 2 of the top-down and bottom-up approaches for building a data warehouse. In the top-down approach, you build the overall corporate-wide data repository using the entity-relationship (E-R) modeling technique. This enterprise-wide data warehouse feeds the departmental data marts that are designed using the dimensional modeling technique. In the bottom-up approach, you build several data marts using the dimensional modeling technique and the collection of these data marts forms the data warehouse environment for your company. Each of these two approaches has its advantages and shortcomings.

You also learned about a practical approach to building a conglomeration of super-marts with conformed and standardized data content. While adopting this approach, first you plan and define the requirements at the corporate level, build the infrastructure for the complete warehouse, and then implement one supermart at a time in a priority sequence. The supermarts are designed using the dimensional modeling technique.

As we have seen. a data warehouse is meant for performing substantial analysis using the available data. The analysis leads to strategic decisions that are the major reasons for building data warehouses in the first place. For performing meaningful analysis, data must be cast in a way suitable for analysis of the values of key indicators over time along business dimensions. Data structures designed using ...