In big data analysis of the bike sharing system (BSS), Tokyo University of Science succeeded in clarifying the statistical characteristics of bicycle usage patterns in real cities.
BSS is attracting attention as a sustainable public transportation system that can reduce CO2 emissions and alleviate traffic congestion.However, while it is convenient for users to rent or return bicycles at any port, the number of bicycles at each port is biased, so an efficient bicycle relocation strategy to eliminate the bias. Is required.
In this study, we believe that it is necessary to clarify the dynamics of BSS of real users and the entire city for such strategic development and social implementation, and we believe that it is necessary to clarify the dynamics of BSS in the four major cities in the United States (Boston, Washington DC, New York, Chicago). We analyzed the usage of bicycles in the actual BSS of.
We selected August, when the number of bicycles used is the highest in the year, and used the historical data of "when and where the user used the bicycle" extracted from all BSS ports in the four cities for analysis.First, it was found that all BSS in the four cities had excess and deficiency of bicycles, and New York had the highest number of excess and deficiency.The frequency of lending and returning was particularly high in New York, but there was a large difference between cities.
Regarding the usage pattern of lending and returning in BSS, a periodic pattern common to the four cities was recognized, such as biased usage in the daytime.Furthermore, similar statistical characteristics were observed on weekdays and weekends, during the day (4:8 to 21:24) and all day (XNUMX hours).However, while New York's bicycle usage pattern is characteristic of the Poisson process (rentals and returns occur randomly without being affected by other events and are not biased per hour), this is not always the case in other cities. It was also discovered that the regularity varies from city to city, suggesting that there is no such thing.
The actual state of BSS in a real metropolitan area found by this study is considered to be an important basis for evaluating and verifying the BSS bias improvement algorithm based on theoretical research using realistic indicators.
Paper information:[Nonlinear Theory and Its Applications, IEICE] Statistical analysis of usage history of bicycle sharing systems
