Abstract:

With the emergence of a large number of globally influential social networking sites like Twitter, the influence of public opinion crosses over the boundary of nation. Network monitoring department began to pay attention to and assess the impact of foreign public opinion on the domestic environment, and Twitter is bound to be a focus of attention. This paper, based on the theory of social media data mining and network analysis, designs and implements a Twitter oriented comprehensive analysis system for data collection and storage, data analysis and interactive display of analysis results with Python, graph-tool, NLTK and other related tools. The system can collect and display the hot topics and related tweets in a certain area in real time, doing sentiment analysis of tweets. At the same time, system can get the relationship network of specified user and specified size, using theory of Closeness Centrality, Betweeness Centrality and PageRank algorithm for individual influence analysis and building Stochastic Block Model for community structure discovery in Networks. System testing is in line with expectations and I hope this system could provide reference and help for the public security work of public opinion on the internet.

Key words: Twitter, sentiment analysis, network individual influence, community discovery, analysis system