Seismic Activity Analysis in Indonesia: Integrating Machine Learning, Geospatial Data, and Environmental Factors for Risk Assessment

Abstract

Earthquake’s phenomena are critical for understanding Earth's interior, tectonic processes, and disaster preparedness.  Because of indonesia location in the Pacific Ring of Fire, it’s suffering from regular seismic activities which result in huge annual losses. This study investigates the earthquake data from 1992 to 2024 by applying clustering techniques such as K-means and geodata visualization. By integrating physics, geospatial analysis, and machine learning, the study processes earthquake data to calculate energy release and analyze spatial-temporal patterns. Principal Component Analysis (PCA) is applied to reduce data dimensionality, while K-Means clustering identifies seismic patterns based on magnitude, depth, and energy. Visual tools, including correlation heatmaps and spatial maps, are used to present findings that support earthquake risk management in Indonesia.The results reveal temporal patterns in earthquake activity, with peaks observed in 2004–2007, associated with significant seismic energy release. Spatial analysis highlights high energy concentrations in megathrust zones. PCA and K-Means clustering identify three distinct clusters with varying correlations between seismic and atmospheric variables, indicating the influence of thermal and tectonic factors. These insights contribute to seismic hazard mapping, risk reduction strategies, and the improvement of earthquake prediction models. Future research should extend datasets and refine machine learning techniques to achieve more accurate predictions.