User Behavior Analysis And Prediction Based On Shared Bicycle Travel Data

In recent years,with the popularity of green travel and environmental protection,sharing bicycles has become a choice for more and more residents.The rapid development of shared bicycles has rapidly declined,which has brought many social problems.How to optimize bicycle resources in the big data environment,improve the service level and make the shared bicycle more benign development through user behavior analysis,become the hot spot that attracts attention and the key to share the competitiveness of bicycle companies.This paper first introduces the research status and trends of shared bicycles at home and abroad,and then introduces the urban public bicycle system and shared bicycles from the concept of shared bicycles,and introduces the related concepts and methods of user behavior analysis.Secondly,aiming at the development status of shared bicycles,it analyzes the current industry characteristics,scale trends and business models of shared bicycles,and the main problems of current shared bicycles,including user quality problems,bicycle maintenance problems,enterprise management and operation issues.Intensive research and analysis were conducted.Then,based on the shared bicycle travel data,the behavior characteristics of the shared bicycle as a whole and individual users are systematically analyzed from various aspects,and the overall travel time,intensity and location of the user are analyzed,and the high frequency users are classified and studied based on the user.Historical travel records,establishing a predictive model for the user's travel destination,and predicting the current date and the user's destination at the current location.Finally,the future development trend of shared bicycles and operational strategies are analyzed to promote the continuous advancement and development of shared bicycles.The results show that there are significant differences in the travel time distribution between the working day and the rest day.The travel intensity has the characteristics of low frequency and short distance.The average number of trips by users is 4-5 times per week,and the travel distance is generally within 3 km.At the same time,by analyzing the riding behaviors on working days and rest days,it is found that the trajectory of high-frequency users sharing bicycles can be divided into three types in the two dimensions of work and leisure,which can help enterprises better understand user travel habits and have targeted Personalized services are provided according to different types of user riding preferences.In addition,the user travel destination prediction model has an average accuracy of about 30%.Since our historical data set is only used for one week,if we increase the amount of historical sample data,it will further improve the prediction accuracy.