Understanding the semantics of data provenance to support active conceptual modeling

Abstract

Data Provenance refers to the lineage of data including its origin, key events that occur over the course of its lifecycle, and other details associated with data creation, processing, and archiving. We believe that tracking provenance enables users to share, discover, and reuse the data, thus streamlining collaborative activities, reducing the possibility of repeating dead ends, and facilitating learning. It also provides a mechanism to transition from static to active conceptual modeling. The primary goal of our research is to investigate the semantics or meaning of data provenance. We describe the W7 model that represents different components of provenance and their relationships to each other. We conceptualize provenance as a combination of seven interconnected elements including "what", "when", "where", "how", "who", "which" and "why". Each of these components may be used to track events that affect data during its lifetime. A homeland security example illustrates how current conceptual models can be extended to embed provenance. © 2008 Springer-Verlag Berlin Heidelberg.