Skip to main content

🧪 AI-Guided SIR Modeling Hackathon (2 Hours)

Welcome to the AI-Guided SIR Modeling Hackathon! In this short, hands-on hackathon, you will learn how to build, simulate, and infer epidemiological models using Python, Google Colab, and ChatGPT.

No prior experience with epidemiological modeling is required.

🕒 Hackathon Format

  • Duration: 2 hours
  • Style: Guided, hands-on hackathon (live demo + participant coding)
  • Tools: Google Colab, Python, ChatGPT / Gemini
  • Setup required: None (no local installation)

All coding, debugging, and explanation will be AI-assisted, with the host demonstrating how ChatGPT can act as a scientific coding co-pilot.

🎯 What You Will Learn

By the end of this hackathon, you will be able to:

  • Understand the SIR (Susceptible–Infected–Recovered) epidemic model
  • Simulate disease outbreaks using ordinary differential equations (ODEs)
  • Estimate key epidemiological parameters such as:
    • Transmission rate (β)
    • Recovery rate (γ)
    • Basic reproduction number (R₀)
  • Fit models to real outbreak data using:
    • Least squares
    • Poisson likelihood
  • Use ChatGPT effectively for:
    • Writing scientific Python code
    • Debugging models
    • Explaining results
    • Generating clean reports

📊 Dataset Used in the Hackathon

We will work with real outbreak data:

  • Disease: Measles
  • Location: Niamey, Niger
  • Source: Grais et al. (2006)
  • Data type: Biweekly measles case counts
  • Communities: 3 (A, B, C)

📌 Important modeling detail:

  • Time is measured in biweeks
  • Average infectious period ≈ 2 weeks
  • This allows us to use a simplified SIR model appropriate for the data

🤖 How ChatGPT Is Used in This Hackathon

ChatGPT will be used as a guided assistant, not as a black box.

You will see how to:

  • Ask clear, precise prompts
  • Specify assumptions (units, data meaning)
  • Request verification (plots, sanity checks)
  • Iterate safely without rewriting everything

🛠 What You Need to Prepare

Required

  • A laptop
  • A Google account (for Google Colab)
  • A modern web browser (Chrome or Firefox recommended)

Not Required

  • No Python installation
  • No prior knowledge of epidemiology
  • No prior experience with ChatGPT

📁 What Will Be Provided

Before or at the start of the hackathon, you will receive:

  • A Google Colab notebook (starter template)
  • The Niamey measles dataset
  • A prompt-engineering cheat sheet
  • All code examples used during the hackathon

You will make your own editable copy of the notebook.

🧭 Hackathon Agenda (High-Level)

  1. Introduction to the hackathon & tools
  2. Brief intro to LLMs and ChatGPT for scientific coding
  3. Estimate R₀ from early outbreak growth
  4. Build and simulate an SIR model
  5. Fit model parameters to data (Least squares method)
  6. Likelihood-based inference (Poisson)
  7. Interpreting uncertainty and results
  8. Wrap-up and next steps

📌 What This Hackathon Is (and Is Not)

✅ This hackathon is:

  • Hands-on
  • Beginner-friendly
  • Focused on learning by doing
  • Guided step-by-step

❌ This hackathon is not:

  • A competitive coding contest
  • A deep theoretical math course
  • A fully automated “push-button” AI demo

🚀 After the Hackathon

You will leave with:

  • A complete, runnable SIR modeling notebook
  • A reusable workflow for epidemic inference
  • Practical experience using ChatGPT for scientific modeling
  • Ideas for extending the model (SEIR, time-varying transmission, Bayesian inference)

We look forward to hacking epidemic models with AI together. See you at the hackathon!