In the digital age, information explosion in cyberspace makes it crucial to extract and analyze valuable information from massive datasets. This project focuses on three main aspects: building vertical search engines and topical crawlers for efficient and precise collection of domain-specific information; conducting sentiment analysis and automatically generating research reports by leveraging natural language processing techniques to gauge textual sentiment tendencies and integrating analytical results into structured reports; and uncovering internet hotspots and interpersonal relationships using clustering algorithms to identify trending topics and the underlying social networks. The project's research findings include public opinion analysis of social hot events, hot event mining, and discovery of special groups. These achievements provide significant insights into the information dissemination patterns and group behavior patterns in complex social networks, offering new perspectives and methods for research and practice in related fields.