hd77alu/eac-climate-risk-index

This project builds a climate‑risk classifier for East African Community (EAC) countries using World Bank climate data. It combines data analysis, preprocessing, and feature engineering with tree‑based models and deep neural networks. The target variable, tx84rr, a temperature‑based mortality indicator, is mapped into three vulnerability classes. # East Africa Community Climate Risk Index This project builds a climate-risk classification workflow for East African Community (EAC) countries using monthly World Bank Climate Change Knowledge Portal data. The notebook combines exploratory data analysis, preprocessing, feature engineering, and two model families: classical tree-based classifiers and deep neural networks. The main prediction target is `tx84rr`, a temperature-based excess mortality indicator that is converted into a three-class vulnerability label. ## Dataset - **Source:** The World Bank Climate Change Knowledge Portal (CCKP) - **Countries code:** BDI, COD, KEN, RWA, SOM, SSD, TZA, UGA - **Collection:** ERA5 0.25-degree (era5-x0.25) - **Type:** Timeseries - **Variables:** `tas, txx, hd30, tr23, hurs, cdd, rx5day, tx84rr` - **Aggregation:** Monthly ## Project Goal The goal of this project is to estimate climate-related public health vulnerability from meteorological indicators and classify each monthly observation into one of three risk levels: - `0` - Normal / Low Risk - `1` - Moderate Risk / Elevated Alert - `2` - Severe Risk / Extreme Health Emergency The notebook focuses on the EAC region and evaluates how well environmental variables can predict periods of elevated health risk. ## Project Structure ```text eac-climate-risk-index/ |-- README.md |-- notebook_mmt_health_risk.ipynb |-- dataset/ |-- cdd-eac-world-bank.csv |-- hd30-eac-world-bank.csv |-- hurs-eac-world-bank.csv |-- rx5day-eac-world-bank.csv |-- tas-eac-world-bank.csv |-- tr23-eac-world-bank.csv |-- tx84rr-eac-world-bank.csv |-- txx-eac-world-bank.csv ``` ### Key Files - `notebook_mmt_health_risk.ipynb` contains the full analysis pipeline, from loading the datasets to model comparison and final interpretation. - `dataset/` stores the monthly World Bank CSV files used in the analysis. ## Data Sources and Variables The notebook uses the following climate indicators: | Category | Variable | Description | | --- | --- | --- | …