Bangladesh has experienced a rapid rise in motorcycle use in recent years. The surge in nonprofessional and professional riders is as a result of demand, ride-sharing opportunities, self-employment, and affordability. Although riding behavior is identified as one of the most influential precursors of motorcycle safety, there is little research comparing the riding behaviors of professional and nonprofessional riders. This study, therefore, aims to differentiate professional and nonprofessional riders based on their distinct riding behavior. Data from 624 motorcycle riders were collected via online and face-to-face questionnaire surveys in Dhaka. Following the feature selection through mean decrease accuracy and mean decrease Gini, this study developed a random forest (RF) model to delve deeper into riders’ behavior. Furthermore, the SHapley Additive exPlanations-based feature was employed to determine the differential factors like overtaking vehicles from the wrong side, carrying passengers without helmets, no speed reduction at intersections, and riding without proper fitness. In addition, this study focused on finding differences in the two rider groups’ law adherence, alertness, and speed behaviors. Consequently, this enables policymakers to design data-driven targeted safety measures that more effectively address the unique risks the two rider groups pose. The findings suggest targeted educational interventions, awareness campaigns, and effective and strict enforcement of the Road Transportation Act of 2018 to improve safety practices. It also recommends purpose-built technology-based solutions, such as the use of wearable smart glasses and AI-integrated speed cameras, as well as engineering-based solutions, such as separating traffic moving at varying speeds and separate lanes for motorcycles.
Sadat et al. (Sat,) studied this question.