Multi-Channel Spatio-Temporal Graph Convolutional Networks for Accurate Micromobility Demand Prediction Integrating Public Transport Data
Pages 5 - 13
Abstract
Accurately predicting city-wide, short-term micromobility (MM) demand, particularly for docked bike-sharing systems, is essential for optimizing urban transportation networks and promoting sustainable mobility. The integration of multimodal data, especially from public transportation (PT) systems like subways, plays a critical role in enhancing the precision of these predictions. However, current data-driven models often overlook the direct integration of public transport checkout data.
To address this gap we propose a novel deep learning approach: the Multi-Channel Spatio-Temporal Graph Convolutional Network (MC-STGCN) with a novel loss function Weighted Normalized Mean Absolute Error (WNMAE). MC-STGCN uses an infrastructure-aware adjacency calculation that factors in cycling and walking times between stations but also provides a learnable adjacency matrix with local and global attention. Additionally, we formalize robustness measures notably the in-city generalizability, essential for adapting to rapid changes and expansions in urban micromobility networks. For high-demand stations, the MM-PT model effectively captures significant morning and evening peaks, At low-demand stations, the MM-PT model significantly outperforms the baseline STGCN, reducing MAE by 20% and handles week days distribution shifts effectively. The MM-PT-W model reduces MAE by 15% compared to GNN-LSTM and 22% compared to ARIMA.
References
[1]
Geoff Boeing. 2024. Modeling and Analyzing Urban Networks and Amenities with OSMnx. https://osmnx.readthedocs.io/en/stable/. https://geoffboeing.com/publications/osmnx-paper/ Working paper.
[2]
Citi Bike NYC. 2024. Citi Bike NYC System Data. https://citibikenyc.com/system-data Accessed: 2024-07-22.
[3]
Antonio Comi and Antonio Polimeni. 2024. Assessing potential sustainability benefits of micromobility: a new data driven approach. European Transport Research Review 16, 1 (2024), 19. https://doi.org/10.1186/sl2544-024-00640-6
[4]
Ezgi Eren and Volkan Emre Uz. 2020. A review on bike-sharing: The factors affecting bike-sharing demand. Sustainable Cities and Society 54 (2020), 101882. https://doi.org/10.1016/j.scs.2019.101882
[5]
Bruno Fernandes, Fábio Silva, Hector Alaiz-Moretón, Paulo Novais, Cesar Analide, and José Neves. 2019. Traffic flow forecasting on data-scarce environments using ARIMA and LSTM networks. In World Conf. on Info. Systems and Technologies. Springer, 273--282.
[6]
Michael Flamm and Vincent Kaufmann. 2006. Operationalising the concept of motility: A qualitative study. Mobilities 1, 2 (2006), 167--189.
[7]
Oleg Golovnin and Nikita Perevozchikov. 2021. E-STGCN: enhanced spatial-temporal graph convolutional network for road traffic forecasting. In 2021 International Conf. on Info. Technology and Nanotechnology (ITNT). IEEE, 1--4.
[8]
Kan Guo, Yongli Hu, Yanfeng Sun, Zhen Sean Qian, Junbin Gao, and Baocai Yin. 2020. An optimized temporal-spatial gated graph convolution network for traffic forecasting. IEEE ITS Magazine 14, 1 (2020), 153--162.
[9]
Yi Hou and Praveen Edara. 2018. Network scale travel time prediction using deep learning. TR Record 2672, 45 (2018), 115--123.
[10]
Jiawei Jiang, Chengkai Han, Wayne Xin Zhao, and Jingyuan Wang. 2023. Pdformer: Propagation delay-aware dynamic long-range transformer for traffic flow prediction. In Proceedings of the AAAI Conf. on AI, Vol. 37. 4365--4373.
[11]
Weiwei Jiang and Jiayun Luo. 2022. Graph neural network for traffic forecasting: A survey. Expert Systems with Applications 207 (2022), 117921. https://doi.org/10.1016/j.eswa.2022.117921
[12]
Andreas Kaltenbrunner, Rodrigo Meza, Jens Grivolla, Joan Codina, and Rafael Banchs. 2010. Urban cycles and mobility patterns: Exploring and predicting trends in a bicycle-based public transport system. Pervasive and Mobile Computing 6, 4 (2010), 455--466. https://doi.org/10.1016/j.pmcj.2010.07.002 Human Behavior in Ubiquitous Environments: Modeling of Human Mobility Patterns.
[13]
Jiyoung Ko and Yung-Cheol Byun. 2023. Analyzing factors affecting micromobility and predicting micro-mobility demand using ensemble voting regressor. Electronics 12, 21 (2023), 4410.
[14]
Yuebing Liang, Guan Huang, and Zhan Zhao. 2022. Joint demand prediction for multimodal systems: A multi-task multi-relational spatiotemporal graph neural network approach. TR part C: emerging technologies 140 (2022), 103731.
[15]
Lei Lin, Zuduo He, and Srinivas Peeta. 2018. Predicting station-level hourly demand in a large-scale bike-sharing network: A graph convolutional neural network approach. TR Part C: Emerging Technologies 97 (2018), 258--276. https://doi.org/10.1016/j.trc.2018.10.011
[16]
Xiaojie Ma, Yuqiang Yin, Yating Jin, Meichun He, and Mingwei Zhu. 2022. Short-Term Prediction of Bike-Sharing Demand Using Multi-Source Data: A Spatial-Temporal Graph Attentional LSTM Approach. Applied Sciences 12, 3 (2022), 1161. https://doi.org/10.3390/app12031161
[17]
Metropolitan transport Authority (MTA). 2024. Subway and Bus Ridership for 2022. https://new.mta.info/agency/new-york-city-transit/subway-bus-ridership-2022 Accessed: 2024-07-22.
[18]
National Association of City transport Officials. 2023. Shared Micromobility 2022. https://nacto.org/2023/11/02/shared-micromobility-2022/. Accessed: 2024-07-15.
[19]
Giulia Oeschger, Páraic Carroll, and Brian Caulfield. 2020. Micromobility and public transport integration: The current state of knowledge. TR Part D: Transport and Environment 89 (2020), 102628. https://doi.org/10.1016/j.trd.2020.102628
[20]
Visual Crossing. 2024. Visual Crossing Weather Data. https://www.visualcrossing.com/ Accessed: 2024-07-22.
[21]
Ke Wang, Changxi Ma, Yihuan Qiao, Xijin Lu, Weining Hao, and Sheng Dong. 2021. A hybrid deep learning model with 1DCNN-LSTM-Attention networks for short-term traffic flow prediction. Physica A: Statistical Mechanics and its Applications 583 (2021), 126293.
[22]
Yu Wang et al. 2022. Cross-mode knowledge adaptation for bike sharing demand prediction. IEEE Transactions on ITS (2022). https://doi.org/10.1109/TITS.2023.3322717
[23]
Yuan Wang, Dongxiang Zhang, Ying Liu, Bo Dai, and Loo Hay Lee. 2019. Enhancing transport systems via deep learning: A survey. TR Part C: Emerging Technologies 99 (2019), 144--163. https://doi.org/10.1016/j.trc.2018.12.004
[24]
Jun-Yuan Xu, Yu Qian, Shi Zhang, and Chun-Cao Wu. 2023. Demand Prediction of Shared Bicycles Based on Graph Convolutional Network-Gated Recurrent Unit-Attention Mechanism. Mathematics 11, 24 (2023), 4994. https://doi.org/10.3390/math11244994
[25]
Hongtai Yang, Rong Zheng, Xuan Li, Jinghai Huo, Linchuan Yang, and Tong Zhu. 2022. Nonlinear and threshold effects of the built environment on e-scooter sharing ridership. Journal of Transport Geography 104 (2022), 103453. https://doi.org/10.1016/j.jtrangeo.2022.103453
[26]
Yang Zhang, Tao Cheng, Yibin Ren, and Kun Xie. 2020. A novel residual graph convolution deep learning model for short-term network-based traffic forecasting. International Journal of Geographical Info. Science 34, 5 (2020), 969--995.
[27]
Qianqian Zhou, Nan Chen, and Siwei Lin. 2022. FASTNN: a deep learning approach for traffic flow prediction considering spatiotemporal features. Sensors 22, 18 (2022), 6921.
[28]
Min Zhu et al. 2023. Improving short-term bike sharing demand forecasting with weather Info. TR Part C: Emerging Technologies 135 (2023), 103984. https://doi.org/10.1016/j.trc.2022.103984
[29]
Weile Zi, Wen Xiong, Hong Chen, and Ling Chen. 2021. TAGCN: Station-level demand prediction for bike-sharing system via a temporal attention graph convolution network. Info. Sciences 543 (2021), 1--16. https://doi.org/10.1016/j.ins.2021.01.065
Index Terms
- Multi-Channel Spatio-Temporal Graph Convolutional Networks for Accurate Micromobility Demand Prediction Integrating Public Transport Data
Recommendations
Long-term drought prediction using deep neural networks based on geospatial weather data
AbstractThe problem of high-quality drought forecasting up to a year in advance is critical for agriculture planning and insurance. Yet, it is still unsolved with reasonable accuracy due to data complexity and aridity stochasticity. We tackle drought ...
Highlights- We improved quality for long-term, up to 12 months drought forecasting.
- We adopted modern transformers and Convolutional LSTM to solve this problem.
- We created an extensive test bed to evaluate models consisting of 5 diverse ...
Combining heterogeneous data sources for spatio-temporal mobility demand forecasting
AbstractThere is a growing need to optimize mobility in medium to large-size cities. The use of a car for one-person trips is widely established as a common trend, which combined with the age of the vehicle fleet from many countries leads to high levels ...
Graphical abstractDisplay Omitted
Highlights- Heterogeneous data is combined for spatio-temporal mobility demand forecasting.
- ST-MDF model forecasts mobility demand for several mobility services and horizons.
- Mobility mesh-grids provide high-resolution spatio-temporal ...
A spatio-temporal graph-guided convolutional LSTM for tropical cyclones precipitation nowcasting
AbstractReal-time and accurate Tropical Cyclone (TC) precipitation nowcasting plays an important role in disaster prevention and mitigation. The strong ability of deep learning has been proved in many precipitation nowcasting studies. However, ...
Highlights- We develop a novel deep-learning framework for TC precipitation nowcasting.
- We ...
Comments
Information & Contributors
Information
Published In
October 2024
38 pages
Copyright © 2024 Owner/Author.
This work is licensed under a Creative Commons Attribution International 4.0 License.
Sponsors
Publisher
Association for Computing Machinery
New York, NY, United States
Publication History
Published: 29 October 2024
Check for updates
Author Tags
Qualifiers
- Research-article
- Research
- Refereed limited
Funding Sources
Conference
SIGSPATIAL '24
Sponsor:
SIGSPATIAL '24: The 32nd ACM International Conference on Advances in Geographic Information Systems
October 29 - November 1, 2024
GA, Atlanta, USA
Acceptance Rates
SUMob '24 Paper Acceptance Rate 5 of 9 submissions, 56%;
Overall Acceptance Rate 5 of 9 submissions, 56%
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 119Total Downloads
- Downloads (Last 12 months)119
- Downloads (Last 6 weeks)52
Reflects downloads up to 25 Jan 2025
Other Metrics
Citations
View Options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in