Recurrent neural networks forecast flu rates across all 50 states up to four weeks ahead.
Hierarchical Bayesian model quantifies the impact of socioeconomic and environmental factors on state-level infection dynamics.