EVALUATION OF TRAFFIC CONGESTION HOTSPOTS USING SMART CARD AND MACHINE LEARNING TECHNIQUES: A CASE STUDY OF IKORODU, LAGOS STATE NIGERIA

ABSTRACT

Traffic congestion in urban cities has become a source of concern as it results in road user inconvenience, greater levels of emission and overall reduction in the quality of life of city dwellers. This research work conducts a comprehensive analysis of smart card transaction data from a Bus Rapid Transit (BRT) system to gain insights into urban transportation congestion. By employing descriptive analytics and Logistic Regression (LR) modelling, a supervised learning modelling technique, travel patterns, congestion levels, and route preferences within the BRT network were investigated. The study evaluated eight BRT routes in order to reveal the most frequently travelled routes, enabling transportation authorities to optimize resource allocation. Congestion levels were categorized into "low", "moderate" and "high" to facilitate effective congestion management. Additionally, a regression model was developed to predict congestion levels, providing a valuable tool for forecasting demand and enhancing operational efficiency. Analysis of the data showed that 50.1% of trips occurred in highly congested areas, 48.8% in moderately congested areas and just 1.1% in low congested routes. It was also discovered that the Ikorodu – TBS route is the most commuted route followed by the Ikorodu – Oshodi and Ikorodu – Igbogbo routes respectively. It also revealed that Tuesdays and Saturdays have the highest number of commuter trips. The performance of the model developed was evaluated using metrics such as recall, precision, accuracy and F1 score. The LR model achieved a precision of 90%, an accuracy of 85%, a recall of 65%, and an F1-score of 69%. This study contributes to data-driven decision-making in the public transportation sector, enabling operational efficiency and passenger experience improvement. In addition, the model presented can be used in various potential applications including improving road safety by warning drivers of upcoming traffic slowdowns and improving mobility through integration with traffic control systems.