, Computational Approaches for Political Redistricting
Computational Approaches for Political Redistricting

This is the webpage for the 2019 MIT IAP course on computational redistricting. Here is the course listing and some additional details. One common theme throughout this course is the idea that ``computational redistricting is NOT a solved problem!''

There are two main ways in which this statement motivates our approach. First, we recognize that there is no one-size-fits-all solution to the challenges of redistricting. Each state has different laws, history, and political geography so any generic approach that optimizes a single measure is doomed to fail. Secondly, and more positively, the fact that there is an enormous amount that we don't know about this proble means that there is plenty of room for creativity in problem solving and algorithm design. There is an air of ``Wild West'' about the current state of the art in both theory and practice and a key goal for this class is to explore the boundaries of this unique problem setting.

Guided by this viewpoint, this course will also attempt to address the following topics:

  1. Understanding how the political redistricting process varies between regions
  2. Understanding how to formalize redistricting as a mathematical problem
  3. Understanding the data challenges presented by combining geospatial information, demographics, and voting data from different (and frequently incompatible) sources
  4. Understanding how to apply MCMC and outlier analysis to the space of permissible plans

Slide Presentations

  1. Introduction and Overview
  2. Geospatial Data
  3. MCMC and Ensemble Generation
  4. Graph Partitioning
  5. Specific State Analyses

Other Resources