Author(s)

BenChao Lan

Affiliation(s)

School of Mathematics and Information Science, Guangxi University, Nanning 530004, Guangxi Zhuang Autonomous Region, China.

Corresponding Author

BenChao Lan

ABSTRACT

To address the uneven spatiotemporal distribution of shared bicycles on university campuses, peak-hour congestion, and insufficient dispatch efficiency, this study targets Guangxi University, aiming to optimize scheduling through demand forecasting. Innovatively integrating univariate and multivariate analyses, it resolves bicycle dispatch challenges while examining student behavioral patterns. Initially, a GAM model revealed that daily rainfall explained 90.08% of ridership variation, with demand exhibiting an exponential decline when precipitation exceeded 8mm. Subsequently, an ARIMA(3,0,0) model confirmed temporal periodicity, and spatial analysis identified academic zones and campus gates as high-demand hotspots. Finally, comparative evaluation of Poisson regression, OLS, and XGBoost multivariate models demonstrated Poisson regression’s superiority for daily predictions, while OLS outperformed in hourly forecasting. Conclusions underscore the strong periodicity and weather sensitivity of campus bicycle demand, affirming that precise forecasting enhances dispatch efficacy. Future work should incorporate variables like class schedules to refine the model, providing a methodological framework for intelligent shared-bicycle management in higher education institutions.

KEYWORDS

Shared bicycles; ARIMA model; Multi-factor analysis; Regression prediction

CITE THIS PAPER

BenChao Lan. Analysis of shared bicycle traffic flow and travel characteristics at a university based on the ARIMA model. World Journal of Information Technology. 2025, 3(4): 80-86. DOI: https://doi.org/10.61784/wjit3059.

REFERENCES

[1] Ministry of Education of the People's Republic of China. Green Travel Development Report in Higher Education Institutions. Beijing: Higher Education Press, 2023.

[2] Liu Jiahao, Wang Chuanyu, Ge Jingxiang, et al. Fine-grained management of campus shared bicycles . Science and Technology Innovation Herald, 2018(5): 176-178.

[3] Tang Yangyang, He Yiyihao, Ji Ning, et al. Analysis of travel behavior characteristics of shared bicycles in university campuses . Traffic & Transportation, 2019, 32(Suppl. 1): 203-206.

[4] Chen Rui, Liu Hao, Wang Yifan. Poisson-GAM hybrid modeling for bike-sharing demand under extreme weather events. Transportation Research Part C: Emerging Technologies, 2023, 45(2): 112-130.

[5] Kim Sungyop, Rudloff Christian, Singhvi Deepak. Temporal periodicity analysis of bike-sharing in educational hubs: A COVID-19 perspective. Transportation Research Part D: Transport and Environment, 2020, 88, 102156.

[6] Li Xuan, Patel Anish, Kim Sungyop. Benchmarking regression models for bike-sharing demand: A university case study. IEEE Access, 2022, 10, 124201-124212.

[7] Patel Anish, Zhang Qing, Chen Rui. XGBoost enhancement for multi-factor bike demand prediction in dense urban campuses. Engineering Applications of Artificial Intelligence, 2024, 129: 107118.

[8] Wang Yifan, Zhang Qing, Zhou Meng. Real-time weather adaptive prediction for shared bikes using sensor fusion. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(1): 89-105.

[9] Wu Jian, Li Xuan, Zhou Meng. Student behavior pattern mining in campus bike-sharing via clustering analysis. Sustainable Cities and Society, 2020, 52: 101189.

[10] Zhang Lei, Li Xuan, Kim Sungyop. ARIMA-LSTM fusion model for short-term bike demand forecasting in university campuses. Expert Systems with Applications, 2021, 185, 115632.

[11] Zhou Meng, Chen Rui, Patel Anish. Peak-hour demand prediction for campus bikes using Fourier transform and ARIMA. Journal of Transport Geography, 2023, 67, 103489.

[12] Wang Lei, Wu Jian, Liu Zheng. Spatial-temporal hotspot identification in university bike-sharing systems via clustering analysis. Proceedings of 2021 IEEE International Conference on Data Science, Virtual Conference. New York: IEEE Press, 2021, 154-201.