Model-driven User Stories for Agile Data Warehouse Development

Agile data warehouse development using Scrum does incremental project delivery by delivering data marts from an initial backlog of user stories. However, story writing is largely ad-hoc and experience-driven. We propose a model driven approach that makes story writing relatively more systematic, and which provides guidance in the story-writing task. The approach is in three layers. At the top-most level, the decision application model contains (a) entities about which decisions are to be taken and (b) Inter-relationships between entities. Criteria for selection of entities are identified. At the next level, decisions for entities are identified. A decision model is proposed and as before, criteria for selection of decisions are presented. Finally, at the third level, information relevant to the selected decisions is modeled and selected to yield user stories. We illustrate the approach with an example of a voice and data communication service provider.