Warehousing, Metadata, and Object-Based Analysis

Warehouse technology amasses teradata as the basis for Decision Support Systems (DSS). Key to knowledge formation is the organization and representation of such data. Identification of data structures enables the selection of decision support models. Use of object-classes with warehouse metadata allows creation of analytic frameworks. We discuss the production application of object-based SAS analyses driven by warehouse metadata. Introduction In the current business climate, competitive advantage requires knowledge-based information delivery. The right information must be readily available for decision makers. Integral to information technology, Data Warehouses organize and maintain a steady stream of information. Welbrock (1998) defines Data Warehousing as “... a process of fulfilling Decision Support enterprise needs through the availability of information” using the key terms: “Decision Support needs” and “availability of information”. In the classical repository view (Sagar and Raval, 1999), warehouses exist primarily to organize input data for DSS. However, business decisions are typically based on condensed analytic results. Interpreting these results requires an understanding of how they were computed (application metadata). Welbrock’s second term implies these metadata definitions must also be warehoused. DSS are often built using object-oriented (OO) methodology, where data is encapsulated in objects. Kolosova and Berestizhevsky (1998) describe a table-driven approach to objectbased SAS programming. Integrating relational warehouse metadata with object processing yields the hybrid architecture shown in Figure 1.