Dataset for: Diagnosing scale-dependent nighttime thermal inequality across heterogeneous city systems using multi-source remote sensing and urban morphology metrics

This replication package contains the processed datasets and Python code necessary to reproduce the key findings and figures of the study on urban nighttime thermal inequality. The analysis integrates multi-source data—including MODIS Land Surface Temperature, ERA5-Land reanalysis data, and building height data—to quantify the impact of 3D urban form on thermal disparity across 24 major Chinese cities. Key components include:1. City-level Metrics: Processed data for two distinct periods (Period 1: 2015–2018; Period 2: 2019–2022), covering Nighttime Thermal Inequality (NTI), built-up intensity, and climatic controls (Ventilation Index, Surface Net Thermal Radiation).2. Causal Inference Code: Python scripts implementing Double Machine Learning (DML) with Random Forest nuisance models. This approach isolates the stable, within-regime effects of built-up intensity on thermal inequality by rigorously controlling for confounding climatic, topographic, and socioeconomic factors.3. Figure Replication: Scripts to reproduce Figure 13 (Mechanism Analysis) and Figure 14 (Temporal Stability of Regime Effects).