Hybrid Data Warehouse Development Method

Abstract

Building a data warehouse is a new discipline and has no concrete strategy for its development process. Currently there are three development approaches for building a data warehouse: Data driven. Goal-driven and User-driven. These development approaches are compared on the basis of certain parameters and by this comparison a new Hybrid multidimensional development methodology has been evolved. This Hybrid multi-dimensional Data model is a combination of Data driven methodology with Business driven which is Goal- driven methodology. We have stated in this paper that this model starts by collecting Business requirements and deriving Fact and Dimension tables along with its multiple constraints which defines their relations. After which a logical structure of the model can be built. Which in turn could be developed into a physical model and can be populated by data for Mining and Analysing. This new multidimensional model can be compared on the same parameters which were used to compare the stated three methodologies and thus we can come up with enhanced features.