Combining Remote Sensing Dataset, Machine Learning, and Explainable AI for Detailed Human Development Index Monitoring in East Java, Indonesia

This study integrates satellite imagery, machine learning, and explainable AI to enhance the predictive accuracy and interpretability of the Human Development Index (HDI) in East Java, Indonesia. Using advanced models, particularly Random Forest Regression, the research achieved an R-squared value of 0.8709, explaining approximately 87% of the variance in HDI. Feature importance analysis identified Night Time Light (NTL) as the most influential predictor with an importance score of 0.4910, indicating a strong correlation with urban development and economic activity. Other significant predictors included the Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI), reflecting the importance of environmental conditions in development metrics. SHAP (SHapley Additive exPlanations) analysis provided insights into individual feature contributions, enhancing model transparency and supporting targeted policy interventions. For instance, higher NTL was associated with increased HDI, highlighting urbanization’s role in development, while pollutants like NO2 and SO2 negatively impacted HDI. This study underscores the effectiveness of integrating satellite data and machine learning in socioeconomic analyses and offers a robust framework for future HDI monitoring.