Automatic data warehouse conceptual design

Abstract

This paper presents a methodology to develop a dimensional data warehouse by integrating all three development approaches such as supply-driven, goal-driven and demand-driven. By having the combination of all three approaches, the final design will ensure that user requirements, company interest and existing source of data are included in the model. We proposed all automatic system using ontology as the knowledge domain. Starting from operational ER-D (Entity Relationship-Diagram), the selection of facts table, verification of terms and consistency checking will utilize domain ontology. The model will also be verified against user and company requirements. Any discrepancy in the final design requires designer and user intervention. The proposed methodology is supported by a prototype using a business data warehouse example.