The research paper investigates the ability of data warehouses projects to meet business requirements. Warehouses are considered to be the complete failed business affairs in most cases. The two different approaches are used to analyze the business objectives: organizational and decisional modeling.
There are three explanations for the projects being unsuccessful. The first one is that the warehousing projects are usually long-termed ones; hence, it is hard to forecast the future business requirements. The second reason is that the information data are communicated poorly across large companies, and ‘jealously guarded by managers’ (Giorgini et al., p.4). The third reason is that the information needed cannot be extracted timely, hence the decisions are taken without the necessary data.
The approaches to data warehousing design are Supply-driven (also called data-driven), and Demand-driven (or requirement-driven). The first approach is efficient and feasible, besides, the data can drawn easily since the information is available.
GRAnD stands for Goal-oriented Requirement Analysis for Data warehouses is implemented by focusing on software engineering that is considered early requirements. GRAnD proposes two perspectives of analysis of requirements for data warehousing: organizational modeling (the foremost importance of which are stakeholders) and decisional modeling (refers to decision makers’ responsibilities).
Interestingly, the data-oriented and the goal-oriented approaches can substitute each other as well as they can work in parallel together to achieve perfectness in design.
As per Tropos in the data warehousing context, it requires some new perspectives and technologies to be applied: facts. Attributes, dimensions, and measures. It is important to mention that the organizational modeling consists of three levels (phases): goal analysis, facts analysis, and attributes analysis. The goal analysis focuses on summarizing the data about stakeholders. It can be presented in various templates, one of which may contain: main actor, sub-actor, and dependencies; then the analysis of each actor’s goals in more detail is carried out. The attributes analysis is the one that gathers the main features that represent importance when facts are being fixed documentarily.
Decisional modeling is an approach that cares much about the role of decision-makers in the company. It is vitally important since it supports the analysis of facts that investigate the decision-maker’s aims.
In the research paper under consideration, the goal-oriented methodology for requirement analysis in DWs is presented. It can be applied successfully to demand-driven and mixed demand-driven design projects. Other approaches offered by contemporary literature are not able to provide efficient techniques to reflect users’ goals to design models.
The discussed methodology was successfully applied to the driving bank’s case study. The user experience with applying the methodology brought fruitful results and helped understand better and validate the approach suggested in the research paper. This can serve as a stable ground to claim that the approach is worth using and will get the highest appraisal of users with high-level goals to design models.
Works Cited
Giorgini, Paolo., Rizzi, Stefano., Garzetti, Maddalena. GRAnD: A goal-oriented approach to requirement analysis in data warehouses. Decision Support Systems 45 (2008) 4–21.