Revealing hidden structure: visualizing bibliographic coupling and co-citation relations in multimedia collections

Many digital collections share a common structure in which a collection, the objects collected and the meaning of the collection can be separately considered. We present a data structure comprising exhibitions, annotations, and resources (EAR) as a general device for organizing such collections. People author EAR structures and other people value these acts of authorship in understanding a large collection. Through co-citation and bibliographic coupling, EAR structures form a general graph that is hard for people to interpret. My research hypothesis is that recognizing, analyzing, prototyping and evaluating the EAR triangle can result in both generalizable insight and new tools for information visualization and system design. I introduce NEAR, a graph visualization tool aimed at helping people understand and use EAR structures. The design, implementation and evaluation follow a user-centered design process in three spiral cycles. The thesis concludes with a discussion of the design and its generalizability.