Traffic counts are crucial for understanding road capacity and ensuring safe and efficient mobility. This study analyzed and predicted traffic conditions on Ikorodu Road, Lagos, using historical daily and hourly traffic data. Machine learning models, including decision trees, gradient boosting, and random forests, were applied. Results showed clear peak and off-peak patterns, emphasizing the need to include factors such as weather, roadwork, and events for more accurate predictions.
This study offers a grounded roadmap for leveraging supervised machine learning—specifically, random forest—to classify and forecast traffic conditions on a busy urban corridor in Lagos. It underscores:
- The importance of rich time-series vehicle count data
- The utility of tree-based ML models for categorical prediction
- The potential for predictive traffic management in real-world contexts
The key takeaway: With stronger data diversity and advanced modeling, Lagos—and other rapidly urbanizing cities—can build intelligent traffic systems that anticipate congestion and deliver smoother, safer mobility.
You can read the paper here