Project Description (Fall only)
The goal of this project is to assess whether a specific state of our choosing is gerrymandered (and if so, to what extent). The main effort in the project will be to design and implement a randomized algorithm to generate a random redistricting of the chosen state. We will then compare the expected voting results under this random redistricting to the actual voting results.
We will use a Markov chain to generate the random redistricting. A Markov chain is a randomized algorithm that does a random walk on thespace of all possible districtings. The challenge is to design a Markov chain that converges quickly (and then to assess convergence). Students should have a good background and strong interest in discrete math and algorithms as we will spend a considerable time studying about the mathematical tools used in Markov Chain Monte Carlo (MCMC) algorithms. Students should also be interested in the social science question of gerrymandering as we will need to find (and clean) appropriate data sets.
The project in this course is intended as 1 quarter to get an initial start on the effort. Interested students who do well can possibly continue for the remainder of the year. For an idea of the possible outcomes here is an undergrad thesis from a Georgia Tech student who was part of a year long project: https://smartech.gatech.edu/handle/1853/63854
- Nicolas Johnson
- Andy Wu
Professor and Mentors
- Prof. Erik Vigoda
- Research group meeting
- Time and Location: Fridays 2PM - 3PM, TBD
Links to Proposals and Presentation
- Proposal (first draft): link
- Proposal (after peer review): link
- Final proposal (after instructor feedback): link
- Final presentation: link