International Science Community Association Research Journal of Engineering Sciences 15 3 2026 09 26 A Data-Agnostic Hybrid SARIMA–LSTM framework for Urban Traffic Monitoring in Smart cities 10 14 EN Amos Omboua Eyandzi 1 Laboratory of Electrical Engineering and Electronic, Polytechnic Superior National School, Marien Ngouabi University, Brazzaville, Congo Flory Lidinga Mobonda 2 Laboratory of Electrical Engineering, ISTA-Kinshasa, Democratic Republic of the Congo Timothée Nsongo 3 Faculty of Sciences and Techniques, Marien Ngouabi University, Brazzaville, Congo André Pasi Bengi Masata 4 Laboratory of Electrical Engineering, ISTA-Kinshasa, Democratic Republic of the Congo Jha Aashish 1 VIVA College, Virar, Palaghar, Maharashtra, India Vohra Shruti 1 Department of Engineering Sciences and Humanity, Chameli Devi Group of Institutions, Indore 452020, MP, India Verma Smita 2 Department of Applied Mathematics and Computational Science, SGSITS, Indore 452003, MP, India Rawat Mohan 3 Department of Engineering Sciences and Humanity, Chameli Devi Group of Institutions, Indore 452020, MP, India > Research Paper 2026 6 9 2026 09 26 The monitoring of urban traffic is critical for effective management in smart city and transport systems. Quite often, existing forecasting approaches tend to be heavily dependent on bulky location-specific data sets, which results in their applicability being questionable under conditions of limited data. In this paper, we suggest developing a novel SARIMA-LSTM approach that combines the principles of statistical time-series analysis with deep learning-based correction of residuals. In the first place, the SARIMA algorithm is introduced in order to guarantee interpretability and statistically proven characteristics of our model due to the identification of linear and seasonal trends. At the same time, the LSTM network is applied to imitate the dynamic non-linearities in residuals in order to improve short-term forecasts. The theoretical development of the approach, its design and experimental verification are provided in the paper. Moreover, the suggested model structure offers an environmentally friendly opportunity for monitoring of urban traffic and facilitates the integration with future spatial-temporal and real-time intelligent transport systems. Copyright@ International Science Community Association