GEPReS: A Geospatially Enabled Predictive Recommendation System for the Preventive Management of Historical Buildings

Abstract

This study introduces GEPReS, a Geospatially Enabled Predictive Recommendation System designed to support the preventive management of historical buildings through short-horizon risk forecasting and context-aware decision support. The system integrates Geographic Information Systems (GISs), Internet of Things (IoT) sensor networks, and authoritative meteorological data to generate timely, actionable recommendations for conservation interventions. These may include pre-emptive shutter closure during heatwaves, activation of ventilation under elevated humidity, or intensified monitoring of structurally sensitive zones during heavy precipitation. By coupling historical datasets with real-time telemetry and calibrated predictive models, GEPReS addresses the distinctive vulnerabilities of heritage structures, which arise from material sensitivity, conservation constraints, and operational limitations under contemporary climatic conditions. The architecture combines spatial analysis, typology-aware risk assessment, and reproducible modelling practices to ensure interpretability and compliance with conservation principles. Designed for scalability and online implementation, the system provides a modular framework capable of adapting to diverse building typologies and resource environments. The paper details the system architecture, data sources, modelling approach, and implementation challenges, supported by empirical evidence from multi-site pilot deployments.