Understanding Corporate Data Models – Shanks & Darke

Shanks, G. and Peta Darke. “Understanding corporate data models.” Information and Management 35 (1999), 19-30.

  • The authors note in the abstract that this article “describes a methodology for designing and representing corporate data models that uses explanation and visualization mechanisms to improve understanding” of data models. Findings include: 1) scenarios are useful for eliciting information requirements and explaining abstract concepts; 2) graphical icons and subject area partitions are effective means of visualizing the model; 3) design rationale is an effective means of explaining the evolution of concepts to data modelers. (19)
  • What is a corporate data model? “an abstract representation of the information requirements of all or part of an organization. [I]ndependent of functional boundaries within an organization and of implementation technology” (19).
  • Data models or data architectures are typically justified on the basis that they “improve the quality of poor or inconsistent data” and help with the generation of more meaningful information vis-à-vis BI (business intelligence) (20). They’re used for futurity by providing a basis for education and communication about information (BI), functioning as a framework for developing “an inventory of data in legacy systems”
  • Conceptual data model: an abstract representation of information requirements (20).
  • The authors argue that explanation and visualization are two key means for improving the viability of data models for important stakeholders throughout an organization (BI, information systems (IS) folks) (20).
  • A major problem in data modeling: the underlying meaning of concepts in the model is misunderstood by the multiple stakeholder groups. As a solution, the authors propose: 1) argumentation-based design rationale; 2) scenario-based analysis; and 3) structured representation of the data model using graphical icons (21). [1. Interestingly, all three of these activities is intensely rhetorical in nature. Argumentation-based design rationale activities involve a negotiation over the contents of a data model and provide a site of consensus building among stakeholders. Scenario-based analysis provide a space for rhetorical analyses, allowing data modelers the opportunity to work with SMEs and users in envisioner and evaluator roles. Visualization of models provide persuasive visual cues/representations that allow stakeholders to understand the overlaps and disjuncts among concepts in a data model. They’re also effective at conveying the entity-relationship model that most data models adapt.]
  • The authors adopt the “ViewCon” methodology in this piece. It is an iterative, two-step process that invites modelers to acquire, model, integrate and test/validate data models. See here:

Figure 1

  • How do modelers initially gather information for the data model? [2. A key area for your own research. This is what we’re proposing to amend with an RGS-Computational Linguistics methodology.] “Specialist data modelers first elicit and accumulate information about the data requirements of the business users from interviews with stakeholders, existing information systems, knowledge of the application domain, and other documentation” (23). [3. An important source: Potts et al. “Recording the Reasons for Design Decisions” Proc. 10th Conf. Software Engineering, pp. 418-427, 1988.]
  • Methods presented for developing the data model: 1) interviews conducted separately and documented using informal narrative description and envisioner scenarios (24); 2) Analysis of the interviews by specialist data modelers; 3) Definition of entity-relationship models for each of the business users; 4) copies of the data models, informal requirements, envisioner scenarios and design rationales were provided to the specialist modelers to integrate the various models into one, more expansive model; 5) Model validation was conducted with graphical models. Validation was evaluated by the business users.

Start the Discussion!Leave a Reply