Hindsight, insight, and foresight: a multi-level structural variation approach to the study of a scientific field

Abstract

In this article we introduce a multi-level structural variation approach to the study of a scientific field. The approach is motivated by an explanatory and computational theory of transformative discovery and novel structural variation metrics derived from the theory. We integrate the theoretical framework with a visual analytic process that enables an analyst to study the literature of a scientific field across multiple levels of aggregation and decomposition, including the field as a whole, specialties, topics and predicates. We demonstrate the use of this approach through a study of regenerative medicine.