Data and Code for "Heatwave risk reshaped by urban expansion-clustering nexus"

This repository contains the core analytical codes and data processing scripts for the study investigating how the urban expansion-clustering nexus reshapes regional heatwave risks. The code is organized into three main Jupyter Notebooks as detailed in the methods section: 1.Urban.ipynb: Contains the scripts for extracting and calculating urban expansion modes, morphological features, and clustering metrics.2.HW.ipynb: Contains the scripts for processing multi-source temperature data, performing bias correction, and calculating the annual heatwave metrics (frequency, duration, amplitude, and onset).3.ML.ipynb: The core causal machine learning pipeline, which incorporates the XGBoost model training, SHAP value interpretation, Accumulated Local Effects (ALE) generation, and the Double Machine Learning (DML) framework. (Note: Data inputs and setup parameters required to run these notebooks are specified within the code comments.)