Towards an Intellectual Atlas of Scholars@Duke
Our goal is to map scholarly production at Duke and use the distribution of scholarship as a frame for displaying relations amongst scholars. Since publications are the fundamental unit of scholarly production, we start by building a publication network and then project a smooth surface-representation of the network in two dimensions in a way that places similar papers close to each other. The distribution of papers across this space represents a landscape with varied topography—hills represent clusters of papers on similar topics, ridges link topics to each other, valleys represent gaps in the knowledge space and widely differing topics are islands disconnected from the mainland. We identify topic content by clustering the network, and label the largest clusters with the terms found most frequently in each cluster’s titles (label size proportional to number of papers). Like a geographic atlas, once we know the topography we can layer other information over this space. Here, we provide layers on high-volume producers, departments/units, gender, and collaboration. The goal of this visualization was to examine the geographic spread of collaborations represented in the scholars@duke dataset. The goal of the Scholars@Duke Visualization Challenge was to create visualizations that capture the richness and dynamism of Duke research. The datasets were provided by Scholars@Duke (https://schoalrs.duke.edu/) and they describe publications, authorships, and scholarly collaborations from university researchers over the past 5 years.
