Multidimensional Data Modeling Based On Multi-subjcets An OLAP Analysis

With the business competition becomes more and more hotly, the MISs couldn't completely meet the leader's requirements. They needed urgently a data warehouse system to help their decision-making. And how to build a data warehouse system based on MISs, how to build a multidimensional data model based on ER model in order to cut the develop circle of data warehouse become more and more important.Data model is one of the key problems of DW. And many researches proved that traditional relation data model couldn't effectively describe the structure and semantic of DW applications, and couldn't meet the requirement of OLAP's applications also. On-line analytical processing is a typical application of DW for the interactive analysis of multidimensional data of varied granularities. Moreover, OLAP supports some multidimensional analytical operations, such as slice,dice,roll-up,drill-down and rotate, which provide users with the flexibility to view data from different perspectives. Because the OLAP is multidimensional, the multidimensional data modeling techniques attract human's general attentions and wide interests.On these situations, this paper mainly does the work as follows:1)introduces the DW and OLAP techniques, concludes the concept and basic structures of multidimensional data model seriously. 2)generalizes the present situation of conceptual modeling for multidimensional data based on ER model, indicates the limitation of applying ER model to OLAP analysis, then formalizes ER model to HPKM, and describes its advantages. 3) Based on the chain store, this paper describes the approach of conceptual multidimensional data modeling based on HPKM, and uses it to model the sales analysis subject, and expatiates how to choose the multidimensional schema according to the requirements of analysis subject. This paper also realizes the classical OLAP queries with multidimensional expended language MDX and visualizes the analysis data.