2 For parking lots of different land-use types, the average number of transactions at different hours of the day, and the percentage of transactions with different parking duration on weekdays and weekends, are shown separately in Figure 2. For example, residential shared parking lots, where the largest number of daily transactions take place, see the first peak time of transactions around 8-9 am on weekdays, followed by an abrupt drop, a second smaller peak time at 18 pm, and a continuous drop until the second day at 6 am. However, on weekends, the number of transactions arrives at its peak around 10 am, then fluctuates until 18 pm before dropping down. On both weekdays and weekends, the largest percentage of parking duration is less than 1 hour. This policy bulletin summarizes a study by Wang, J., & Zhu, P. (2024) that uses data from 121 shared parking lots in Guangzhou, China to analyze the influence of a set of explanatory factors on shared parking use, addressing the following questions: • What are the spatial distribution features of the shared parking lots? When do the shared parking transactions start in a day and how long do they last? • How do the features of shared parking lots and urban spaces influence the total number of transactions? How do they influence the average parking duration? Study Methodology The study uses 418,635 transaction records data of 121 shared parking lots in Guangzhou, China from November 2020 to October 2021, which are derived from the Chinese largest shared parking platform – Airparking. Each shared parking lot record contains the information of parking lot name, coordinates, land use, construction type, implemented duration, and capacity. Each transaction record contains the information of parking lot name, car type, reservation placing time, waiting time, parking start time, parking end time, timeout duration, transaction grading, and open comments for the transaction. Moreover, geodata of Guangzhou, including the information of POIs (points of interest, e.g., tourism attractions, office buildings, and shopping centers/streets), are extracted from online open source data. Based on the data, the study first discusses the spatial and temporal distribution features of shared parking. After that, a quasi-Poisson regression model is introduced to understand influential factors on the total number of transactions at different shared parking lots, while a linear regression model is introduced to understand influential factors on average parking duration. Findings and Analysis Spatiotemporal Distribution Features of Shared Parking Figure 1a to 1d show the spatial distribution of shared parking lot features. As shown in Figure 1a, 78.51% of shared parking lots are clustered in city central districts, i.e., Tianhe District (35.54%), Yuexiu District (21.49%), and Haizhu District (21.49%), which fits the higher parking demands in city central districts arising from higher urbanization rates and population density. Moreover, residential shared parking lots are most widely distributed, early in every district, while business office shared parking lots mainly cluster in central administrative districts. Most shared parking lots with more than 500 transactions in the observed year also cluster in central administrative districts. Location of shared parking lots in the city (left) and central districts (right) Figure 1a Land use type of shared parking lots in the city (left) and central districts (right) Figure 1b Implemented duration of shared parking lots in the city (left) and central districts (right) Figure 1c Factors Influencing Shared Parking Use in Time and Space: A Case Study in Guangzhou Public Policy BULLETIN
RkJQdWJsaXNoZXIy NDk5Njg=