2020-Election-Prediction
Predicting the Outcome of the 2020 Election: k-NN, logistic regression, LASSO/ridge regularization, model fairness evaluation
Description of notebooks:
ac209a_election_prediction_analysis.ipynb
: Contains the majority of our project work, including data processing, EDA, modeling, and forecasting for 2020 Presidential and House of Representative races.ac209a_election_prediction_covid.ipynb
: Contains our analysis of unemployment data, which we used as a proxy for COVID-19 effects, and includes a number of helpful visualizations included in our report.
Description of data:
acs_2013_variables.csv
: state-level demographics dataacs_pop_density_2010.csv
: state-level population densityelec_college_votes.csv
: 2020 state-level electoral college voting outcomeshouse_results_76_18.csv
: district-level congressional election outcomes from 1976-2018potus_results_76_16.csv
: state-level presidential election outcomes from 1976-2018president_polls.csv
: state-level polling data from 1972-2016state_name_crosswalk.csv
: state names, shortened names, and 2-letter abbreviations, used for matching across data sourcesurbanicity_index.csv
: state-level population and urbanicity data