Election Forecasting Documentation

State-level presidential election forecasting using polling time-series data from the 2016 U.S. presidential election.

Contents:

Features

  • Four different forecasting models (Poll Average, Kalman Diffusion, Improved Kalman, Hierarchical Bayes)

  • Comprehensive evaluation metrics (Brier Score, Log Loss, MAE)

  • State-level time-series visualization

  • Automated model comparison pipeline

Installation

pip install election-forecasting

Or with uv:

uv pip install election-forecasting

Quick Start

# Run complete pipeline: forecast, compare, and plot
election-run-all

# Run individual commands
election-forecast
election-compare
election-plot

Indices and tables