A CS333 Project for MCMC-algorithmic redistricting for fairer congressional maps.

This notebook is the main technical demo for this project. It is a Jupyter notebook that imports Colorado data, represents its counties as an undirected graph, and iterates through the MCMC algorithm to produce a new congressional map of Colorado designed to generated roughly-equally-populated congressional districts that are politically competitive.

This Python files contains various helper files for the MCMC algorithm used in TechnicalDemo.ipynb. It should be included in the same folder as TechnicalDemo.ipynb to not cause any ModuleNotFound errors.

Cleaned CSV file that takes in all [County, Neighbor] pairs in county_adjacency.txt, such that both counties are in Colorado (CO). If there exists a pair of counties [a, b] in the CSV, the pair [b, a] will also exist in the CSV.

CSV file compiling various statistics from partisan lean, population, and registered voters.

Data Dictionary:

• County: County name (in Colorado, sorted in alphabetical order)
• PVI: Partisan Voting Index
  • Represents whether a county favors the Democratic or Republican presidential candidate more
  • Positive values indicate that the county favors the Democratic candidate
  • Negative vaues indicate that the county favors the Republican candidate
  • Zero represents a neutral county
• Population: County population based on 2022 estimates
• #Democrats: Number of Democrats in the county as of 2023 estimates
• #Republicans: Number of Republicans in the county as of 2023 estimates
• White: Percentage of white Coloradans in the county as of 2023 estimates

This Python file contains helper functions that reads from county.csv and data_by_county.csv to return an undirected graph representing Colorado. Each node has attributes for name, population, and PVI. Edges are added based on contiguity of counties.

All files contained in the scrape/ folder are related to data scraping, collecting, cleaning, and formatting. This folder is how we extracted data for Colorado counties and used it for the algorithm.

Map displaying all of Colorado's 64 counties, along with their neighboring counties. Source: https://ccionline.org/info-center-library/maps/

Adjacent county data sourced from the US Census at https://www.census.gov/programs-surveys/geography/library/reference/county-adjacency-file.html [1]

Lists all counties and their adjacent neighbors in "all 50 states, the District of Columbia, Puerto Rico and the Island Areas (American Samoa, the Commonwealth of the Northern Mariana Islands, Guam, and the U.S. Virgin Islands)." [1]

When the data was collected, the page was last revised on December 16, 2021.

Python file used to clean county_adjacency.txt into county.csv. Ran on Python 3.11.2 and requires the re and csv libraries to run.

Finds the white percentage of a county's population for each county in Colorado. Source: https://demography.dola.colorado.gov/assets/html/population.html

Finds the partisan lean (PVI) of each county in Colorado. Source: https://www.zipdatamaps.com/counties/state/politics/map-of-partisan-voting-index-for-counties-in-colorado

Finds the population of every county in Colorado. Source: https://gis.dola.colorado.gov/population/population-totals-counties/

Contains Jupyter notebooks used to experiment with code prior to creating the technical demo notebook.

Jupyter notebook that demonstrates how to extract node and edge information from county.csv to be used for NetworkX graphs.

Contains plots of MCMC algorithm scores over iterations for evaluation purposes.

Folder used to store images over iterations of the MCMC algorithm for video creation.