A Simple Abstraction for Data Modeling

Abstract: The problems that scientists face in creating well designed databases intersect with the concerns of data curation. Entity-relationship modeling and its variants have been the basis of most relational data modeling for decades. However, these abstractions and the relational model itself are intricate and have proved not to be very accessible among scientists with limited resources for data management. This paper explores one aspect of relational data models, the meaning of foreign key relationships. We observe that a foreign key produces a table relationship that generally references either an entity or repeating attributes. This paper proposes constructing foreign keys based on these two cases, and suggests that the method promotes intuitive data modeling and normalization.