Algorithmic Democracy: Mathematical and Supercomputer Approaches to Gerrymandering

Authors: John Hessler*, Library of Congress
Topics: Legal Geography, Geographic Information Science and Systems, Geographic Theory
Keywords: Gerrymandering, GIS, redistricting, supercomputing, markov processes, evolutionary algorithms,
Session Type: Paper
Day: 4/5/2019
Start / End Time: 1:10 PM / 2:50 PM
Room: Maryland A, Marriott, Lobby Level
Presentation File: No File Uploaded


Gerrymandering has a long history in law, politics and cartography. Today however, with the use of specialized algorithms and supercomputers, it has become a mapping project very different from what it was in the 19th and 20th centuries. This paper will give an introduction to the modern science of gerrymandering and highlight how massively parallel computation is giving rise to new forms of cartography and geospatial information based on the processing of huge amounts of census and social media derived data. Unbelievably, these simulations are creating billions of maps that reveal hidden patterns in voting behavior, and have led to new and interesting forms of analysis and visualization that have spawned deep questions concerning what constitutes a gerrymandered map. The paper will provide an understanding of evolutionary algorithms, Markov Chain models and review current mathematical research programs that use Ricci curvature, exploring real case studies using the PEAR algorithms developed at the Blue Waters National Supercomputing Center.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login