Analysis Of Rainfall Impact On Transit Ridership

Beijing,as a typical metropolis,also serving as the capital of China,carries the core functions of "political center,cultural center,science and technology innovation center,international communication center" for the whole country,and shoulders the task of building a comprehensive,integrated transportation system that is safe,convenient,efficient,green and economical.However,during the peak period of high throughput,the traffic system is very fragile,and the occurrence of highly influential weather conditions will lead to the rapid deterioration of service capability and operational efficiency,causing large-scale and long-term congestion.Rainfall is a kind of highly influential weather condition.In order to improve the overall transit efficiency,the traffic planning department grants great priority to the construction of public transport infrastructure and the development of public transport.Public transport,because of its convenience,punctuality,welfare,is of the wide favor among residents.In order to improve the ability of transportation system,so as to cope with the impact of high influential weather,and provide a scientific basis for transportation decisionmaking in rainy weather,it is necessary to analyze the rainfall influential regularity on public transportation in Beijing.The smart card data(SCD)collected from the intelligent equipment provided a good data base to mine the potential regularity of public transit behavior.However,there is a lack of appropriate methods to effectively analyze the rainfall impact in-depth.Under this backdrop,based on the large SCD,this paper explored the basic laws of the residents' travel,excavated the characteristics of residents' public transportation through a variety of clustering algorithms,confirmed the significance of rainfall impact on public transport trips,proposed an impact analyzing method based on the ridership fluctuation coefficient,and mined the different rainfall impact regularities on the ridership of different characteristics on the minute timescale.Specifically,the main content and contribution of the paper can be summarized as the following three aspects:(1)In order to deal with the big data of SCD efficiently,and grasp the basic statistical information of public transportation,this paper carried out data preprocessing through the distributed computing framework “MapReduce”,calculated the subway and bus ridership with an interval of 5 minutes,and analyzed the basic statistics of cyclicality,tidality,travel time distribution,travel distance distribution and travel frequency distribution of public transportation.(2)In order to analyze the background,purpose and the complexity of public transit travel,this paper operated several unregulated clustering learning method on the data of pre-processed rail traffic card: achieve the feature extraction of site passenger flow pattern through K-Means algorithm,achieve the travel elasticity feature extraction through DBSCAN algorithm,and finally achieve the trip feature extraction through LDA algorithm.This provided a solid basis for further analysis of different rainfall impact.(3)In order to analyze how various meteorological factors influence the transit ridership,the multiple linear regression model was adopted and it proved the significance of rainfall impact.In order to analyze the process of rainfall impact quantitatively,this paper calculated the base ridership with the ARIMA model,put forward the analysis method based on “fluctuation coefficient ridership”,and analyzed the different situations of how ridership of different characteristics get influenced differently.This paper mined the basic regularity of public transit,and put forward the analysis method of rainfall influence on minute timescale,and deeply analyzed the process and degree of rainfall impact on public transportation.Several conclusions can be drawn as the following: rainfall had a significant impact on public transit,rainfall impact is closely related to the passengers characteristics.Long distance trips and flexible trips were more vulnerable to rainfall,while trips flowing to airports or tourist attractions could be sharply cut down due to rainfall impact.These conclusions provided scientific basis for the traffic department to make management decisions,and for the public transportation operators to optimize service and promote the humanization of public transport.