Hacking the Election

This repository represents all the code written for the research project Hacking the Election: Measuring and Solving Gerrymandering in Today’s Political System. It includes Python and C++ for parsing data, detecting Communities of Interest, measuring gerrymandering, and an algorithm for redistricting.

More information is available on the website at hacking-the-election.github.io, along with an interactive tool for displaying our results.

---

Installing Data

All of our code is ran from raw data collected from the following sources:

• @nvkelso/election-geodata
• @unitedstates/districts
• @mgg-states
• Open Precincts
• Harvard Election Data Archives
• California Election Database

Our data is split into directories by state, each containing a .zip with some combination of precinct geodata, election data, district geodata, and population data.

To install only our selection of data and not our code, the standalone data repository can be cloned:

git clone 'https://github.com/hacking-the-election/data'

This includes all of the raw data, as well as binary state files for each language. This data repository is additionally included in this gerrymandering repository as a git submodule. To clone and install:

git clone --recursive https://github.com/hacking-the-election/gerrymandering

---

C++

Currently, the output for this project was run using the C++ implementation of our algorithm. It is being developed into a library with a much better API than it has currently. However, currently our programs and binaries can be installed as shown below.

Installation

git clone "https://github.com/hacking-the-election/gerrymandering"
cd gerrymandering/cpp
make all
make install  # May need sudo privileges

This will compile all available programs. If you want a selection, the following binaries are available and can be installed from the gerrymandering/cpp directory with make <binary_name> && make install.

• serialize_state <geodata.json> <election_data.tab> <district.json> --keys='{keys}' <output.state>: to be used in conjunction with build/serialize.sh - used to build state binary files. These state binaries are essential for all following programs.
• generate_communities <infile.state> output/directory: Run the communities algorithm on a certain state, outputting images and quantification of current districts.
• state_dump <infile.state>: dump information on a certian state file

OS Compatability

• All code will be compatible with macOS and Linux
• Binary distributions will be available after completion of all code

---

Python

This code will only work with a python version >=3.8.0

Installation

git clone "https://github.com/hacking-the-election/gerrymandering"
cd gerrymandering/python
pip install -e .
python3 setup.py build_ext --inplace  # Compile the Cython.

There is, however, one dependency which is not on PyPI, and must be installed by cloning its repo.

If you would like to install without pip, you know what you're doing, or you really, really don't know what you're doing. All dependencies are listed in python/requirements.txt

This package will be availalbe on PyPI sometime in the future.

Testing

To run all unit tests, run the command:

python3 -m hacking_the_election test

OS Compatability

• Any software written in Python will be Windows, macOS, and Linux compatible.