Dengue Forecasting Hackathon 2025

Small Island Nations, Machine Learning Team

Data dictionaries: https://github.com/kraemer-lab/global_dengue_forecasting/blob/main/docs/README.md

Pipeline

Target time-series and covariates are preprocessed using notebooks/Preprocessing_climate_data_hackton_8dec.ipynb. This produces /data/Covariates_data_{iso}.csv files for each country ISO code.

Usage

Run in a Python virtual environment, we recommend uv. This code was developed with Python 3.13.2.

uv venv
uv pip install -e .
source .venv/bin/activate

Then, launch a forecast, e.g. for the example NHiTS model:

python scripts/nhits.py