Keynote Talk: Business intelligence: Dimensional modeling fundamentals and challenges

Abstract

Dimensional modeling is a methodology for the design of data warehouse systems for decision support and analytical requirements, as an alternative to traditional online transactional processing systems. The fundamental steps, as well as the common techniques and good practices of dimensional modeling are presented, including granular (star schema) design for flexibility, conforming dimensions for large system integration, slowly changing dimensions, and real-time partitions to support near real-time data warehouses. Data warehouse systems face more challenges than traditional (transactional) systems. These challenges could be functional or technical. Some of these are addressed, including issues with respect to data structure (or lack of structure), (big) size, integration, and (front end) analytical applications complexity.