Learning Bayesian Statistics #130 The Real-World Impact of Epidemiological Models, with Adam Kucharski
Apr 16, 2025
Adam Kucharski, a professor of infectious disease epidemiology, dives into the art of epidemic modeling and its vital role during crises like COVID-19. He discusses the challenges of communicating complex models to the public and the importance of Bayesian statistics in navigating uncertainty. The conversation also explores how ideas and diseases spread similarly, emphasizing the need for collaborative efforts in public health. Plus, Kucharski reflects on the impact of AI in improving data interpretation and decision-making in epidemiology.
AI Snips
Chapters
Books
Transcript
Episode notes
Data Are Often Modeled Estimates
- Many data called 'raw' in epidemiology are actually model-derived estimates.
- Epidemiological models often evaluate scenarios rather than precise forecasts.
Write Down and Update Predictions
- Encourage politicians and journalists to write down and update their predictions.
- This practice improves probabilistic thinking and understanding of data.
Bayesian Spread Beyond Epidemics
- Epidemic thinking informs fields like finance and marketing through concepts like reproduction numbers.
- Belief updating in people often follows Bayesian principles rather than backfire effects.
