Getting Started

Install python dependencies

pip install -r requirements.txt

Enable plotly plots in jupyter lab

bash enable_plotly.sh

Generate Polling Statistics

See generate_sample_size_figures.ipynb

Evaluate Predictions

See evaluate_preds.ipynb

Figures

The figures directory contains all generated figures visualizing polling data, prediction accuracy, and more.

figures/data

In the figures directory, the data directory holds all the csv files that were used in generating the figures.

Data Directory

election_data - Results from the NYTimes 2020 general election data, as well as processed "final results" (the most recent values for each state).

label_data - This directory contains a few different files.

• Normalized vote share results for 2016 presidential race
• Normalized vote share results for 2018 senate races.
• Vote share results with battleground classification for 2016 presidential election.
• Electoral vote data, to be used for final election outcome prediction.

poll_data - Poll data contains all polling data for the presidential race. It also contains cleaned, normalized, vote share labels from the raw poll data.

population_data - Population data, generated from FRED.

predictions - Predicted noramlized vote share for each model.