Research On Quasi Real Time Customer Behavior Model Of Data Warehouse System Of DaZhou Bank

Da Zhou bank is a city commercial bank in Sichuan Province.It not only has the inevitability of development,but also has the limitation of development.It is the product of market,but also has the limitation of region.Compared with the large and medium-sized commercial banks at home and abroad,it has developed for decades late.It is in the initial stage and also in the stage of vigorous development both in technology and business.In the process of development,many technical and operational problems have been produced.This paper takes the data warehouse system of Da Zhou bank as the research object,and expounds how to solve some problems encountered in the development process of Da Zhou bank.The main contents of this paper are as follows:(1)This paper studies the method of batch data automatic matching collection of sharkstart tool.Based on the point-to-point manual configuration collection of sharkstart tool,this paper studies the collection method of sharkstart tool's batch data configuration,automatic matching collection method and the preprocessing method of illegal characters,so as to reduce the human intervention,shorten the development cycle and improve the development efficiency.(2)This paper studies the subject model of quasi real time customer behavior based on traditional data warehouse system of bank.Using Hadoop big data technology,including Kafka real-time data collection technology,stream real-time data analysis technology,Map Reduce real-time data calculation technology,HDFS real-time data storage technology,real-time processing of customer behavior data.Combined with the static data subject model of the traditional data warehouse system,this paper studies the quasi real-time customer behavior subject model,realizes the dynamic and static data subject model of the bank data warehouse system,and improves the comprehensive use value of business data.(3)Research data visualization technology.Based on the data visualization of static table form of moiavisual tools,this paper studies the integration of Yonghong Bi visualization development tools that can be dynamic graphics in moiavisual tools,so as to improve the display effect of quasi real-time data visualization.Based on the above research content,the main research results and innovations are as follows:(1)This paper proposes a quasi real-time customer behavior subject model based on bank data warehouse system.According to the needs of customer behavior analysis,based on the traditional static data model,a subject analysis model combined with the dynamic customer data of data warehouse is designed,and the quasi real-time customer behavior analysis is preliminarily realized,which provides a more practical means for customer behavior analysis.(2)This paper presents a method of data batch automatic matching collection based on sharkstart.On the basis of sharkstart's data collection function,excel template configuration is used,Python and SQL technology are used,the data collection mode of the original system is improved,the batch automatic data collection function is realized,and the efficiency of data collection is improved.To sum up,through the construction of data warehouse system of Da Zhou bank,the technical problems in ETL process are solved;the technical problems in data visualization process are solved;the subject model design of quasi real-time customer behavior is completed;the quasi real-time data visualization application of executive cockpit is realized;the static data visualization application of comprehensive report and supervision report in the bank is realized It has improved the market competitiveness,prevented the business risks and met the supervision requirements of the people's Bank of China.