| With the rapid development of the Internet,e-commerce,Internet banking and Internet education have draw many users ’ attention because of their convenience and price advantage.Taobao,the most representative e-commerce website in China,has nearly 500 million registered users.Such a huge user base will inevitably generate massive data about users,commodities and transactions.In the era of DT(Data Technology),how to collect,store and use the massive data,mining user’s behavior patterns and potential demand,have important significance for both users and sites.This thesis summarizes the background and research status of e-commerce websites at first,and then discusses the Hadoop distributed data processing platform,which provides a platform support and technical support for the storage and calculation of massive data.Analysis of traffic and user behavior mainly have three aspects,including sites,users and mobile e-commerce.First,we mainly study business penetration and traffic model,and describe the website operation conversion rate funnel model for optimizing website design to enhance the user’s purchase.Second,we mainly explore the differences between user scale and traffic on different terminal types and operating systems to set up a personalized customers and marketing management model in mobile Internet.Third,through establishing user profiling and tags level library,we help sites understand users’ habits and needs to promote precise recommendation.Based on real Internet data,this thesis provides e-commerce websites analysis and comparison of traffic and user behavior patterns from multiple perspectives,which have significant implications for mining and using the value of massive data. |
