TY - JOUR
T1 - Framework for hourly demand forecasting of bike-sharing stations
T2 - case study of the four main gate areas in Seoul
AU - Hong, Jungyeol
AU - Han, Eunryong
AU - Park, Dongjoo
N1 - Publisher Copyright:
© 2024 The Institute of Urban Sciences.
PY - 2024
Y1 - 2024
N2 - Shared bicycles represent a sharing economy for solving complex urban traffic problems. Therefore, their demand has been steadily increasing since the introduction of shared bicycles in Seoul. The demand for shared bicycles is influenced not only by temporal characteristics but also by various factors such as the characteristics of the city, the environment around shared bicycle rental station, and physical urban network. Therefore, the primary purpose of this study is to discover the factors affecting the demand for shared bicycles and develop models that predict the demand for each shared bicycle rental station over time, reflecting the influence of these factors. In this study, 263 shared bicycle rental stations in the four main gates at the centre of Seoul were classified through time-series clustering analysis, and the demand of each rental station was estimated by time using the random forest method. Consequently, it was found that the amount of rental and return an hour before and the temperature and precipitation an hour before were significant factors in predicting the demand for the next period. Furthermore, it was found that the cluster model considering the characteristics of time-series changes was more accurate than the models that were not cluster-specific. It is expected that future research will monitor the inventory of bicycles at rental stations and establish strategies for relocation using the predicted demand obtained by the framework of the analysis.
AB - Shared bicycles represent a sharing economy for solving complex urban traffic problems. Therefore, their demand has been steadily increasing since the introduction of shared bicycles in Seoul. The demand for shared bicycles is influenced not only by temporal characteristics but also by various factors such as the characteristics of the city, the environment around shared bicycle rental station, and physical urban network. Therefore, the primary purpose of this study is to discover the factors affecting the demand for shared bicycles and develop models that predict the demand for each shared bicycle rental station over time, reflecting the influence of these factors. In this study, 263 shared bicycle rental stations in the four main gates at the centre of Seoul were classified through time-series clustering analysis, and the demand of each rental station was estimated by time using the random forest method. Consequently, it was found that the amount of rental and return an hour before and the temperature and precipitation an hour before were significant factors in predicting the demand for the next period. Furthermore, it was found that the cluster model considering the characteristics of time-series changes was more accurate than the models that were not cluster-specific. It is expected that future research will monitor the inventory of bicycles at rental stations and establish strategies for relocation using the predicted demand obtained by the framework of the analysis.
KW - Bike-sharing
KW - demand forecasting
KW - random forest
KW - temperature
KW - time-series clustering
KW - urban network
UR - http://www.scopus.com/inward/record.url?scp=85185672366&partnerID=8YFLogxK
U2 - 10.1080/12265934.2024.2317198
DO - 10.1080/12265934.2024.2317198
M3 - Article
AN - SCOPUS:85185672366
SN - 1226-5934
VL - 28
SP - 735
EP - 750
JO - International Journal of Urban Sciences
JF - International Journal of Urban Sciences
IS - 4
ER -