Evaluating Usability and Public Engagement of MRT Network in Dhaka: A Machine Learning Approach

Abstract: This study will develop a machine learning framework aimed at evaluating and forecasting public interaction trends within the MRT transportation network (MRT Line-6) located in Dhaka, Bangladesh. This research aims to optimize service usability and inform strategic decision-making for sustained future user engagement. The project seeks to analyze public interaction and usage patterns recorded from April 2023 to November 2024 using the Daily Origin-Destination (OD) report from the Dhaka Mass Transit Company Limited (DMTCL). This data will be used to build a predictive model capable of accurately projecting future trends. The methodology will proceed in structured stages: data preprocessing will standardize the dataset, followed by feature selection and variable definition to isolate critical factors influencing user engagement. A robust predictive model will be developed using advanced machine learning algorithms, with an 80/20 split between training and prediction datasets to ensure high model accuracy. The study will incorporate cross-validation and iterative refinement to enhance reliability. Besides, expert consultations and pilot testing with MRT professionals will validate the model to ensure practical applicability and accuracy, identifying opportunities for system enhancements. The resulting model will provide actionable insights into public behavior patterns and usage fluctuations, supporting data-driven optimizations for MRT service delivery. This approach not only strengthens transportation management practices for the MRT network in Dhaka but also establishes a replicable framework that can be adapted to improve public transportation systems in other urban settings, thus contributing to the broader field of transportation analytics.