The ticketing management of railway department usually makes decision based on historical data or empirical analysis of experts. That leads to the decision-making much vulnerable and fluctuate，then effects the development of railway. With the development of the network and information technology，railway has also driven by the big data. The characteristic of the big data is configurable rather than single. So it becomes very important to railway section to invest how to use the big data to serve the railway development and help them to make significant decision. At the same time，the data mining in railway mainly focus on the construction of database. Nevertheless，the research on the data analysis method are rare in the railway ticketing management. There are just several papers about how to apply the train ticket data into the railway marketing. With the motivation of better applying the big data into the railway ticketing marketing，this thesis has the fowling aspects:
(1) According to the relevant analysis and reduction method to select the reference data which is based on the character of the date. And using analysis of variance significant correlation with test data and found that there is significant correlation between booking time and other factors.
(2) Analysis the correlation of the advance time of booking ticket with personality character, train type, buy ticket mode, the type of the destination city and the travel time.
(3) Using density-based clustering algorithm DBSCAN gets railway tickets on classification category 5 customer groups, explaining the result of the five categories and each type of customers’ characteristics are analyzed.
(4) At last，finding the distribution regularity of railway passenger ticket behavior, many significant suggestions are proposed based on the derived of data analysis and data mining.
谷歌
火狐
