Case study on navigating data elements by faceted classifications in scientific data warehouse

In a traditional data warehouse, the “subject oriented” means that all the elements (data attributes) relating to the same real-world event or object are linked together, the elements are classified and navigated by “subject”. But in scientific data warehouse, elements are “faceted classifications oriented”, users want to classify and navigate the elements by mutable classifications flexibly. It is hard to find a universal “subject” to support all usages. This paper is a case study of such issues, we propose a classification cube whose cells are modeled as cross-classification specified data marts. It supports users to navigate to any data marts easily by the query and OLAP operations, further provides the effective data service to the front-end applications.