Traffic accidents pose the most significant challenges to public safety and urban development in metropolitan areas like Manila. This study uses a methodology based on Geographic Information Systems (GIS) in conjunction with machine learning, applying the K-Nearest Neighbors (KNN) algorithm, to predict the traffic accident hotspots. It integrates geographic and contextual features to create risk maps that identify accident-prone areas using spatial data and historical accident records. The predictive model had an accuracy of 97% with high precision but a little lower recall, thus providing opportunities for enhancement. GIS-based visualizations provide actionable insights that are helpful in traffic management strategies, resource allocation, and urban planning efforts. The approach therefore points to the potential of spatial analytics integration with machine learning for better road safety.