Early information requirements engineering for target driven data warehouse development

Abstract

We propose two stages for data warehouse requirements engineering (i) an 'early information' part where the information relevant to decision making is discovered, and (ii) a 'late' part where this information is structured as facts and dimensions. Our focus is on the former. Early information data warehouse requirements engineering starts with targets defined as pairs of the form where A is an aspect of an organization and I is a set of business indicators. An aspect is a work area, work unit, service or quality to be preserved in an organization. Business indicators are measures/metrics for specifying the desired performance level of aspects. Targets are organized in a target hierarchy. This hierarchy is a complete specification of what is to be achieved by a top level target. We associate targets with choice sets so that alternative ways of target achievement can be represented. These alternatives form their own hierarchy. Finally, information relevant to selection of each alternative is discovered through Ends, Means, Key Success Factor, and Outcome Feedback analysis techniques. These techniques determine early information that is to be subsequently to be processed in the 'late information' requirements engineering stage. Our early information requirements engineering phase is illustrated through a case study. © IFIP International Federation for Information Processing 2012.