Data warehousing commonly implies complex data flows, either because of the large number of steps data transformations may consist of, or of the different types of data they carry. These issues rise interesting challenges concerning design-oriented modeling of data warehousing flows with UML.
This article identifies and formally defines data warehousing flows (fw’s) as the founding concept for every visual modeling technique studied so far, which means that flows can be applied to model any data transformation, from OLAP to data mining. It also presents an integrated visual modeling framework for fw’s based on UML. The proposal consists in the diagramming of data cubes, meant as results of multidimensional queries and data transformations. These diagrams have been implemented in a prototype. This solution covers the modeling gaps identified in the state-of-the-art of data warehousing flows.
Read the complete article on http://www-db.deis.unibo.it/~srizzi/PDF/dawak09.pdf