Hybrid Tree Visualizations for Analysis of Gerrymandering

Chenguang Xu, Sarah M. Brown, Christan Grant, Chris Weaver

Abstract: In the United States, congressional redistricting follows a decennial Census. Gerrymandering can result from selection of district lines regardless of political parties. Understanding the relationships between the multiple dimensions in electoral data is a core goal of gerrymandering analysis. In this paper, we analyze patterns of gerrymandering in election data using a hybrid tree visualization technique that supports both overview and drill-down into a hierarchy of multidimensional relationships in that data. Visualization of hierarchical data is of major interest in information visualization. The technique utilizes a left-to-right node-link diagram to show overall hierarchical structure. Nodes in the diagram depict the levels in the dimensional hierarchy. Each node is rendered as an embedded view that shows its particular dimensional combination. Edges directly connect the contents of the embedded views to provide visual bridges that aid navigation and understanding of dimensional relationships. We demonstrate the utility of this hybrid technique is demonstrated through two use cases. This work aims to both ground and inspire the design of future visualizations for exploring gerrymandering.