Research On Urban Public Bicycle System Sites Demand Prediction And Data Analysis

Public bicycle is a healthy and environmentally friendly way of transportation,which provides convenience for people’s travel.However,the unreasonable distribution of sites,the imbalance of bicycle supply and demand at every site in the peak period,the imbalance of operation and management are in the process of running the public bicycle system,which restricts the development of the public bicycle system.The key to solve these problems is to analysis user behavior,operate sites,predicte the bicycle demand in the future period.This paper chooses the public bicycle system in San Francisco Bay Area as the research object,and studies the user behavior characteristics and the operation rules of the site.The main contents of this paper are as follows:(1)The background of the study of the public bicycle system and the significance of using large data technology to solve the existing problems of public bicycles are described.At the same time,the research status of the public bicycle system at home and abroad is introduced.(2)Combined with the historical data of the public bicycle system in the bay area of San Francisco,data processing is done by using big data tools such as Spark SQL and HDFS.In addition,user behavior characteristics and site operation rules are also studied,and visual analysis is realized by means of Highchart,Origin,Python and other tools.(3)A big data computing platform based on Hadoop and Spark is built.Based on it,a prediction model of public bicycle demand based on random forest is built.In the prediction,the user data,site data and weather data are used as input variables,and the demand of the site is the output variable.The demand of the bicycle at different stations in the next day is predicted,and the prediction results are analyzed and verified.(4)The quantile regression strategy is introduced into the prediction of bicycle demand,and the RBF neural network with nonlinear ability is used in the quantile regression to realize the interval estimation of the demand of bicycles at different stations in the next day.At the same time,the relationship of each factor to the demand of the bicycle is fully revealed.