Data-driven Cycling Quality Evaluation And Bike-metro Transfer Demand Forecast

With the development of society,urban traffic management has gradually improved from "rough" governance to "refined" governance,and the concept of urban traffic governance is people-oriented and low-carbon traffic.Encouraging bicycle travel has become an important way to help "carbon reduction" and "carbon peaks".Objectively and truthfully evaluating the conditions of cycling facilities and calculating the potential demand for cycling under urban stock renewal are important for improving cycling environment.How to take advantage of the convenience of the bicycle in short-distance travel,increase the proportion of bike-metro transfer mode in medium & long-distance travel,improve the overall efficiency of cycling,and reduce carbon emissions,which has become a key issue of government departments and researchers.Whatmore,the cycling quality evaluation and bike-metro transfer demand forecast are two core tasks for improving riding environment.With the advent of the "Internet +" era,the emergence of shared bicycles such as OFO and Mobike has provided good services to the citizens,and at the same time,it has also laid a data foundation for reflecting the actual bicycle traffic demand and supply.However,compared with vehicle data,the shared bicycle positioning data from mobile phones’ GPS is noisy and difficult to process,which seriously restricts the application of shared bicycle big data in bicycle traffic management.Therefore,in view of the practical need for refined and high-quality governance,this research uses the big data and data modeling and analysis technology of shared bicycles in the existing cycling network to focus on the actual cycling state under the interaction between supply and demand of bicycle traffic,and proposes a research framework to realize the iterative improvement of the cycling environment in a gradual update way.In order to solve the problems of cycling data processing,cycling behavior feature extraction,cycling quality evaluation,and connection cycling demand estimation,the data processing and trajectory extraction technology of shared bicycles based on mobile phone GPS were deeply explored,and the bicycle riding quality evaluation method and the potential connection demand prediction method for rail passenger flow were constructed,which laid basic theoretical support for the improvement of the riding environment under the existing stock.On this basis,combined with the actual situation,the process method that meets the work requirements is proposed.The research contents are as follows:(1)Cleaning strategy and preprocessing of shared bicycle dataThis chapter analyzed the shared bike data content composition,distribution,noise,and riding behavior characteristics presented in the shared bike data.The article proposed a data cleaning strategy and preprocessing method,which can support the study of quantitative measurement methods for riding behavior and refined evaluation of cycling environment quality.Considering the impact of shared bike data noise on the data quality and availability,this research proposed a limited correction data processing strategy with a certain noise tolerance and gave a pseudo-riding record cleaning processing method and a method for identifying and processing abnormal trajectory based on box graph.This method can provide reliable basic data support for the extraction of cycling trajectory information and model construction.(2)Cycling track extraction based on HMM modelThe chapter analyzed the characteristics of mobile positioning data and the cycling behavior of shared bike users.and studied the method of riding trajectory with the HMM model road network that is compatible with it.Unknown,lacking direction and speed attributes,large positioning deviations,and more suspension groups have adverse effects on the accuracy of space data matching.The implementation of the matching model has been improved,and the matching data sequence of shared bicycle positioning data sequences to the sequence sequences of the road network sequence sequences will be achieved,and the logical relationship between riding positioning data and road network space will be constructed.Provides the foundation of riding trajectory data after the space information is provided.(3)Riding environment quality evaluation and the problem identification methodBased on the demand for refined development of bicycle traffic management,the chapter constructed a multi-scale evaluation system for cycling process quality,which is based on cycling behavior characteristics.An evaluation system has used of the individual riding behavior characteristics contained in the shared bike and the "point segment" as a slightly basic evaluation unit.Based on the data characteristics of shared bikes,a quantitative measurement method for cycling behavior characteristics is proposed.On this basis,this research considered the characteristics of the data,as well as the needs of the systematic spiral improvement work that gradually iterates the weak points of the riding environment,used data mining methods and data analysis techniques,proposed a riding quality evaluation method with the characteristics of data-driven and multi-model integration.(4)The bike-metro transfer demand forecast methodThis research has taken the bike-metro transfer of shared bicycles at the peak as research objects.Based on land use data,distribution data of station connection passengers,shared bike trajectory data,and road section riding quality data,the study constructed a distance attenuation model of rail stations,which considered the influence of factors such as the selection of station connection methods,the spatio-temporal characteristics of passenger flow attraction at rail stations,and the distribution of occupational and residential population in the attraction area of rail stations.Using the connection sharing probability model of the riding distance threshold,a method of the demand forecast model for bike-metro transit is proposed,which is based on the spatial characteristics and selection behavior at rail stations.(5)The research of a bicycle system project workflow and an exampleTaking the existing bicycle traffic system in Beijing as the research object,the riding situation under the interaction of supply and demand was analyzed by using the shared bicycle data.The working method of improving the quality of the cycling environment was proposed based on the requirements of intensive,high-quality and progressive promotion mode.Taking the sixth district of the city as an example,this research evaluates the riding quality,identifies the problem road sections,and takes the promotion of the rail transit service level of "rail + bicycle" as the starting point,selects the Renmin University Station in the Huangzhuang area of Haidian to predict the potential demand for connection,and analyzes the surrounding situation and puts forward the transformation and improvement strategy based on the results of cycling quality evaluation,identification of problem road sections and prediction of potential demand for connection.The results of this research can enrich the existing research methods in the field of urban cycling environment improvement theory and bicycle travel behavior under urban stock renewal,and provide important theoretical and methodological support for policymakers to formulate policies and strategies for urban cycling environment improvement.80figures,15 tables,149references.