Comprehensive Transportation Management System with Driver Performance Evaluation Using Random Forest Algorithm

Several public transportation services in the Philippines, such as the Transport Service Multi-Purpose Cooperative, still rely on paper records and basic tools. This leads to difficulties with data management, transit operation, and monitoring driver performance. To address these issues, this study developed a Comprehensive Transportation Management System with Driver Performance Evaluation to improve operational efficiency and service quality. Web technologies were employed to provide functionalities such as driver management, vehicle management, scheduling, dispatching, finance management, reporting, feedback systems, data visualization and driver portal. Additionally, a machine-learning Random Forest algorithm was integrated for evaluating driver performance, achieving 91% accuracy. Following the implementation of the system, an assessment was conducted using the ISO 25010 quality model, with feedback collected from 10 administrators and 70 drivers. The results demonstrated strong performance across key quality attributes. Functional Suitability received a score of 4.67, classified as “Very Satisfied,” while Performance Efficiency followed closely with a mean score of 4.65, which was also rated as “Very Satisfied.” Reliability achieved the highest satisfaction with a mean score of 4.75, maintaining the “Very Satisfied” classification. While still positive, Security received a lower score of 4.49, classified as “Satisfied.” Overall, the system scored 4.64 out of 5.00, reflecting an overall rating of “Very Satisfied.” The results imply that the study introduced significant innovations to the Pandacan Transport Service Multi-Purpose Cooperative.