Data Quality By Design: A Goal-Oriented Approach

We present a design process for incorporating data quality requirements into database schemas that is rooted in goal-oriented requirements analysis techniques, developed in the Requirement Engineering community over last 15 years. Goal-oriented approaches (i) offer a body of notations, techniques and processes for modeling, analyzing and operationalizing quality requirements; (ii) support representation and evaluation of alternatives in goal fulfillment; and (iii) provide automated reasoning tools for various analysis and design tasks. This paper extends existing proposals for addressing data/information quality issues during database requirements analysis and design in two ways. First, we consider a broader range of quality assurance data requirements, which can be classified as restrictive, descriptive, supportive and reflective. Second, we offer a systematic way to operationalize high-level, abstract quality goals into operational, concrete quality assurance data requirements plus standard operating procedures, based on a risk-based analysis. In an earlier paper, we presented a goal-oriented conceptual database design approach, focusing on deriving an ordinary conceptual schema from application-specific goals. In this paper, we take the next step to incorporate quality goals into the design process. The proposed quality design process is illustrated step-by-step using a meeting expense database example. We also show how this process fits into the overall goal-oriented conceptual database design approach to offer an integrated framework for analyzing both application-specific and quality assurance data requirements.