Geospatial and Data-Driven Risk Management Methods for Tracking Anomalies in Transportation and Environmental Systems

This dissertation presents a unified framework for proactive risk analysis and systems management across critical lifeline infrastructures, specifically for the domains of transportation safety and environmental resilience. With advances in connected vehicle (CV) technologies, machine learning, and geospatial analytics, this research introduces scalable, data-driven methodologies to identify, predict, and mitigate risks in both human mobility and agricultural systems. There are three component methods as follows. (i) The first component of this research addresses pedestrian safety in high-risk school zones by applying clustering techniques (DBSCAN) and hotspot analysis (Getis-Ord Gi*) to CV data capturing harsh braking and acceleration events. These techniques enable proactive identification of hazardous driving patterns before crashes occur, supporting Vision Zero initiatives and equitable infrastructure planning. (ii) The second component explores the information management systems required to use such insights at scale. Through the development and implementation of a geospatial, model-based application, this study demonstrates how digital twins and centralized knowledge repositories can support data governance, multi-criteria decision-making, and lifecycle infrastructure planning. (iii) The third component expands the geospatial risk modeling framework to environmental systems, focusing on the impact of extreme heat on agricultural productivity. Using machine learning models (XGBoost, SVR, Random Forest) alongside spatial interpolation methods (Kriging and IDW), the study generates heat-based risk indices to assess vulnerability in food systems and human health under changing climate conditions. With the above synthesis of geospatial anomalies, or hot spots, methodologies across transportation and environmental sectors, this dissertation thus contributes a comprehensive, systems approach to risk management with real-time data, spatial intelligence, and predictive analytics. The results support scalable, cross-sector solutions that enhance resilience and safety in the face of evolving threats to infrastructure and public well-being.