How can we leverage newly developed computational tools to gain insights into gerrymandering and related redistricting problems?

How can we democratize access to these tools to empower diverse stakeholders to participate in their state or community’s redistricting process more effectively?


Creating truly “neutral” maps that are fair, representative, and unbiased is a difficul goal. Computational methods have expanded greatly since the 2010 redistricting cycle. The 2020 redistricting cycle is the first time that computational tools are widely available for researchers, redistricting commission members, state legislators, advocacy groups, and community members. There is great promise in the power these compuational tools have to offer in regard to the many challenging problems posed by congressional redistricting. However, computational redistricting is not a solved problem. This work lies at the intersection of multiple fields – mathematics, political science, public policy, human geography, history, legal studies – and requires expansive and in-depth study to advance towards generally agreed upon processes and applications. Our project aims to contribute to redistricting work by evaluating one newly developed computational tool, GerryChain, and using these insights to develop a guide that will help community members, activists, and non-partisan redistricting committtees access these tools more easily.



Several states passed crucial reforms to the redistricting process in 2018 in an effort to curb gerrymandering practices. These reforms have expanded the legal framework that state legislators and independent election commissioners must adhere to while creating new districting plans. However, to ensure that proposed plans are aligned with these legislative reforms, analytic methods must be put in place. More broadly, careful analysis is needed to ensure proposed plans uphold principles of fairness and democratic representation. Recent computational and statistical advances provide tools to help map drawers evaluate potential tradeoffs between redistricting criteria in the map generation process. Additionally, once a plan is proposed, relevant metrics can be used to assess the fairness of plans.

Gerrymandering has been a persistent problem in the US since 1812, but new developments in computation have provided a novel set of tools for addressing this issue. Academic groups at Princeton, Duke, and Tufts have been working in this space for the past several years, developing the underlying mathematical theories and computational tools necessary to assess gerrymandering of legislative maps. Members of these groups have also testified as expert witnesses in suits brought against legislative maps. In addition, they’ve also supported non-partisan map drawers attempting to rectify maps deemed unconstitutional, and in one instance, even produced an amicus curiae brief submitted to the Supreme Court in Rucho v. Common Cause (2019); supporting the application of these methods for assessing legislative maps on the basis of partisan or racial bias. However, 2021 is the first redistricting cycle in which maps are being drawn while these ensemble-based analysis tools are publicly available. This means that the current redistricting cycle is a novel opportunity for map drawing commissions, consultant groups, and advocacy organizations to apply these tools to produce fairer maps or to evaluate and challenge potentially gerrymandered legislative maps as they are developed.



We aim to uphold principles of fairness and democratic representation in the redistricting process. We consider fairness both in terms of process and outcomes of redistricting. We seek to maximize inclusion and transparency in the redistricting process and our guide is largely motivated by the goal of democratizing access to resources, computational tools, and knowledge necessary to particiapte in redistricting. In outcome, we view fair maps as those structured for equal representation of voters regardless of identity or partisan leaning to align with the principle of one person, one vote.

We acknowledge that in any nationwide study, such as ours, there’s a risk of flattening local and state-conditions and the need to honor communities. We give consideration to the legacy of racial segregation, redlining, racialized geographies, and the history of race relations and exclusion in this country. We understand that redistricting often reinforces these conditions. We want to prioritizing the Voting Rights Act in our analysis, and roleplay bad intent/exploitation for these tools while designing a guideline for accessibility.

GerryChain as a computational tool for redistricting is open source and widely available. Yet there are considerable technical barriers to its uptake and usage. In an effort to expand the pool of potential users, we have focused on creating annotated case studies applying GerryChain in select state contexts and developing a comprehensive user’s guide. Our goal is to enable anyone with modest technical skills to utilize the GerryChain toolset for the purpose of analyzing or challenging potentially gerrymandered maps in their own states. We engaged with a variety of stakeholders to understand their perspectives and roles in the redistricting process, as well as to capture greater nuance about their specific state and local contexts.