The Analysis System Of E-commerce Real-time User Behavior Based On HADOOP

With the development of Internet and information technology,"online and offline combined industry springing up like in all walks of life to flourish development,all line industry have begun to on-line.In the background of the increasingly fierce competition,enterprises can seize the market opportunity,better development.How to find a better breakthrough,is eager to solve the problem.Especially under the Internet data explosion type of growth.Produce a large amount of data every day.Data has become a resource.It is as valuable as gold and energy.Big data has become another way of thinking of the enterprise.How to mine the valuable information from these massive data,become the important target of the enterprise.Big data brings new opportunities to enterprises,but also brings new challenges,massive data storage,collection,calculation,analysis is the need to solve the problem.This paper analyzes the main mode of user behavior analysis in the context of domestic and foreign companies,including the relevant theory and process,and put forward a set of real-time analysis system of user behavior,which is based on the rapid development of e-commerce.So that it can take advantage of the characteristics of the high efficiency for the enterprise to seize the initiative,to bring effective value,to achieve business growth.Real time user behavior analysis system based on Hadoop and storm open source framework,using Java technology to develop.The system is divided into three modules: data acquisition layer,data storage layer,and data analysis display layer.The subsystems of each module are coordinated to each other,and realize the real-time analysis of the website user data.The system has the following features: 1)the use of Hadoop distributed storage to deal with massive data storage;2)flow computing engine can analyze the behavior of web users,and generate real-time results.3)through the Web front-end technology,multi dimensional visualization data display.At present,the system has been implemented in a Agel Ecommerce Ltd,and the data access,real-time data acquisition,real-time data processing,and the user behavior analysis model based on the data processing.To the final production of accurate data decision statements.Can provide a good support to the decision-making layer.This system can be easily extended to the level,the ability to increase the ability of data storage and parallel computing.Is considered to be a good performance,high reliability,strong expansibility,the development framework is simple system.Not limited to this,the design idea of the module of the module can let the system can be extended to a wider area.Such as real-time monitoring,real-time warning,etc..