REVIEW OF THE DATA MODELING STANDARDS AND DATA MODEL TRANSFORMATION TECHNIQUES

Manual data transformations that result in high error rates are a big problem in complex integration and data warehouse projects, resulting in poor quality of data and delays in deployment to production. Automation of data transformations can be easily verified by humans; the ability to learn from past decisions allows the creation of metadata that can be leveraged in future mappings. Significant improvement of the quality of data transformations can be achieved, when at least one of the models used in transformation is already analyzed and understood. Over recent decades, particular industries have defined data models that are widely adopted in commercial and open source solutions. Those models (often industry standards, accepted by ISO or other organizations) can be leveraged to increase reuse in integration projects resulting in a) lower project costs and b) faster delivery to production. The goal of this article is to provide a comprehensive review of the practical applications of standardization of data formats. Using use cases from the Financial Services Industry as examples, the author tries to identify the motivations and common elements of particular data formats, and how they can be leveraged in order to automate process of data transformations between the models.