Preference Considered Network Data Visualization

With the rapid development of internet in recent years, it has made considerable progress both in terms of internet scale and content. Internet applications such as e-commerce and social network have become the fastest growth areas. It is a big difference between traditional retail business and social network and the activities on internet, traditional research method has a difficulty to study the law of users on the internet. Researches done on the most important of the three network user behavior, online shopping, social networking and network information sharing show that network behavior has obvious attribute of aggregation of the preferences of user behavior.This thesis is based on several performances of network user behavior, particularly the data involved in hot topics of the forum users, the popular microblogging messages and Taobao transaction data, it has done researches of cluster analysis, and preferences based visualization considering the emotion of the user, user attribute, the topic property, commodity recommend and seller’s credit. For forum data processing, it employed text analysis combined with astrolabe considering user participation, visualize the group and individual preferences effectively. Astrolabe is employed in this thesis to visualize the macro preferences based on user’s comment and retweet, it shows the guiding role of VIP users. For Taobao transactional behavior, it set up a recommendation model based on commodity comment, seller’s credit and the correlation between different kind of commodities. Finally visually analyze the distribution law of the commodity and recommend possibly buying goods for users considering user preferences.In this thesis, three different types of data were the establishment of the visual model, it put forward ideas from a new angle to internet user behavior visualization which avoid ignorance of user preferences.