Forecasting Urban Fire Severity for Enhanced Emergency Response and Resource Allocation

The paper is to develop a predictive model to help fire departments improve resource allocation by estimating the likelihood of fire escalation and using GIS to enable faster data-driven decision-making, thereby improving efficiency and public safety. The study used fire data from 47,382 incidents to employ an XGBoost model that predicts fire severity based on building characteristics. After cleaning and preprocessing, the model was trained and validated with 5-fold cross-validation and tested across various contexts. The study found that building characteristics, such as structure, use, and floors, strongly predict fire severity. Fires are more likely to escalate in older buildings, at night, and on weekends. With 82.7% accuracy, the predictive model could reduce property damage, injuries, and response times through better resource allocation. This paper introduces a novel use of predictive analytics for firefighting resource allocation. It applies an XGBoost model to predict fire severity based on building characteristics. Unlike prior research, it focuses on real-time decision-making and resource deployment. The one-year trial shows the model’s potential to enhance firefighting efficiency, offering new insights into how building design affects fire escalation and improving public safety strategies.