Design And Implementation Of User Portrait System Based On Microblog Data

With the rapid development of the Internet and breakthroughs of Internet users,enterprises have accumulated a large amount of user data. In today’s era of big data,enterprises are unwilling to simply save the user data, but hope to tap the business model which has not been found using data analysis techniques. But the traditional method relies upon personal experience heavily, so we need a simple but powerful platform to make it easy.The User Portrait System Based on Microblog Data uses clustering algorithms such as K-means to analysis the users, then for each group of users, analyze it and give it tags.After the analysis, the user becomes a collection of short text which can describe the user well. The system can help us to understand the user deeply. Every short tag of the user is based on the analysis of large amounts of data. In this way the user portrait system improves the availability of data greatly and helps us to get out from the complex user data.The system provides a rich user interface, not only provides a page inquiry services, but also provides protocol inquiry services. In page query section, the system uses the mainstream industry LAMP structures. Protocol inquiry service is divided into three-tier architecture, interface layer, logic layer and data layer. Each layer consists of a number of servers and supports dynamic growth and reduce.The system provides a platform to analyze the user and makes it easy using variety of data analysis methods and data visualization methods. Besides, the query module uses heartbeat package, cache and standby machine to ensure the system stable and scalable.