Advisers

Khristian M. Hoseña

Discipline

IT & management

Abstract

Local Government Units (LGUs) face challenges in maintaining community infrastructure due to reactive maintenance practices and limited asset condition data. Roads, drainage systems, bridges, and other facilities are often repaired only after visible damage or public complaints. This approach increases costs, causes service delays, misuses resources, and creates safety risks. This study proposed a Community Infrastructure Maintenance Management System for LGU 1 that integrates predictive analytics using the Azure AutoML Regression Model. The system forecasts maintenance needs, estimates costs, and supports preventive scheduling using historical maintenance and environmental data. It promotes data-driven decision-making and proactive infrastructure management. Developed using Agile Scrum, the system is scalable and flexible. It enables LGU 1 to shift from reactive to predictive and preventive maintenance, improving infrastructure lifespan and service delivery. A developmental and evaluation-based design was applied. The system was built through Agile Scrum with iterative feedback. Core modules included asset monitoring, maintenance scheduling, predictive analytics, work order management, and reporting dashboards. Predictive modeling used Azure AutoML Regression to automatically select the best model from historical data such as repair records, costs, environmental factors, and asset usage. System evaluation combined quantitative analysis of forecast accuracy and qualitative feedback from LGU personnel to assess usability and effectiveness. The developed system was reported to support maintenance planning, resource allocation, budgeting, transparency, and task management. Predictive estimates were described as reliable, and emergency repairs were described as reduced through early identification of high-risk assets. Predictive analytics may support preventive infrastructure management when historical data are complete and users are trained. The system demonstrates a proposed data-driven workflow using Azure AutoML, but claims that it improved efficiency, reduced costs, extended asset life, or transformed service delivery require quantitative validation and operational deployment evidence.

Keywords

predictive analytics, community infrastructure maintenance, azure automl, regression model, preventive maintenance