Abstract by Annika King
Gerrymandering in Utah: Markov-Chain Monte Carlo
Various measures of Gerrymandering are used across the U.S. However, each state is unique in its geography and voter population, so it’s naive to believe that the same score means the same thing across all states. Using Markov-chain Monte Carlo methods, we constructed a large ensemble of alternative district plans for the state of Utah to give a qualitative aspect to the scores Utah received. We discuss the methods used to create new districting plans as well as methods and algorithms to implement in the future, including parallelization and genetic algorithms. Results of comparing the 2012 districts to an ensemble of plans are also discussed in depth.
Note: This presentation should follow Gerrymandering in Utah: Relevance and Leveraging Data by Cristina Lange and precede Gerrymandering in Utah: Markov-Chain Monte Carlo Methods with Gibbs Sampling by Jacob Murri