Study On Special Travel Behaviors Using Subway Smart Card Data

The travel behavior of passengers can reflect the basic situation of a city.However,extreme conditions often draw our attention and point to important underlying mechanism,such as the most depressed city of a region,the most popular gateway city among immigrants and so on.Affected by the global financial crisis,the number of the unemployed,self-employed,part-time workers,and telecommuters has increased over the past decade.These special groups can expose the city's problems to a certain extent.As the wide application of public transportation automatic billing system,smart card with physical space-time tags has provided a new method for methodology for studying passengers' special travel behavior.Based on the multi-source data,such as the one-week subway smart card in Wuhan from March 23 to March 29,2015,land use data,house price data,and on the premise of mastering the travel characteristics and classification theory of passengers,this paper adopts data mining,clustering,regression analysis,data visualization and other technical methods to analyze special travel behavior of passengers from two directions of "human" and "place".First,to understand the general travel characteristics of passengers,this paper investigates the characteristics of the spatial and temporal distribution of subway passenger flow based on the full sample;Second,from the perspective of "people",the paper provides working definitions of special travel behavior and guesses possible corresponding groups;Then from the perspective of "place",according to the special travel characteristics index,the subway stations are classified to identify the distribution of abnormal stations;Finally,with the help of land use data and house price data,the paper seeks to find the underlying causes behind these special behaviors.The result shows that:1.The traffic volume of Line 2 is the largest among the three lines,and the stations with more traffic volume are mainly located near the commercial districts;2.The flow and flow direction show a high symmetry whether from group scale or loop scale;3.One passenger may have several special travel behaviors;4.Different from the distribution of the full sample,the special traffic volume of Line 1 is the largest among the three lines,and Line 4' is the fewest;5.These special groups may be residents far away from the city center,serviceman,inspector,and staff members working for the subway.They may also be part-time flyers,hackers,beggars,tramps and so on;6.The distribution of early birds and tireless itinerants sites are positively correlated with logistics warehousing land,confirming the speculation that these groups may be the employees of logistics companies.The distribution of night owls sites is positively correlated with commercial commerce,entertainment and recreation land,confirming the speculation that overtime or recreational activitie leads passengers to go home late.All sites are positively correlated with entertainment and recreation land.