PECO Gas Leak Predictor

Gas leaks pose serious safety, environmental, and financial risks for utility companies such as PECO, the project stakeholder. Currently, most leak management strategies are reactive, resulting in high emergency-response costs and infrastructure downtime. This project aims to develop a web application with a user-friendly interface and a machine learning backend model that predicts potential gas leaks before they occur. By analyzing historical leak data, environmental factors, and pipeline characteristics, the system will identify high-risk areas and recommend proactive maintenance actions. The application will allow stakeholders to visualize leak-risk predictions, track system performance, and make data-driven repair or replacement decisions. A predictive approach will reduce emergency repairs, improve safety, and enhance operational efficiency.