Stop-Based Time Series Traffic Change Analysis Using Data from Istanbul Metropolitan Municipality (IMM)

Abstract

Traffic congestion is a critical urban issue that affects travel efficiency, air quality and public well-being. This study examines traffic density patterns in the coastal region of Beşiktaş¸, Istanbul, by analysing hourly vehicle counts from September 2024. The research identifies peak congestion during the morning (07:00-09:00) and evening (16:00-19:00) rush hours, with higher traffic density on weekdays compared to weekends. Ordinary least squares regression shows a weak inverse relationship between traffic density and average speed, highlighting the need for additional variables to increase explanatory power. Three predictive models - Seasonal ARIMA (SARIMA), Facebook Prophet and Random Forest - are evaluated for predictive accuracy. The results suggest that Random Forest provides superior short-term forecasting accuracy, while SARIMA and Prophet effectively capture seasonal trends. These findings provide a robust framework for urban traffic forecasting and management, supporting the development of informed strategies to reduce congestion in dense urban areas.

Keywords

• Time series
• SARIMA
• Random Forest

References

1. Ghosh, S., Basu, B., O’Mahony, M. (2007). Bayesian Time-Series Model for Short-Term Traffic Flow Forecasting. ASCE Journal of Transportation Engineering, *133*(3), 180–189.
2. Han, X, Shi, X. (2015). Online Traffic Congestion Predicton Based on Random Forest. Proceedings of the IEEE International Conference on Intelligent Transportation Systems, 102–107.
3. Hong, W. (2011). Traffic Flow Forecasting by Seasonal SVR with Chaotic Simulated Annealing Algorithm. International Journal of Computational Intelligence Systems, *4*(4), 568–576.
4. Kumar, P., Vanajakshi, L. (2015). Short-Term Traffic Flow Prediction Using Seasonal ARIMA Model. IEEE Intelligent Transportation Systems Magazine, *7*(2), 45–55.
5. Liu, H., Wu, J. (2017). Prediction of Road Traffic Congestion Based on Random Forest. IEEE Transactions on Intelligent Transportation Systems, *18*(2), 377–385.
6. Wang, X., Liu, Y. (2018). Traffic Volume Prediction on Busy Road Junctions. Transportation Research Part C: Emerging Technologies, *95*, 21–36.
7. Zarei, M., Zarei, R., Sattari, A. (2013). Road Traffic Prediction Using Context-Aware Random Forest Based on Volatility Nature of Traffic Flows. Proceedings of the IEEE International Conference on Intelligent Transportation Systems, 175–180.
8. İstanbul Metropolitan Municipality. (2024). Hourly Traffic Density Data Set https://data.ibb.gov.tr/dataset/hourly-traffic-density-data-set/resource/914cb0b9-d941-4408-98eb-f378519c26f4