A Data Analysis Framework for a Topic on Social Media

This study aims to systematically analyze data on a specific topic from social media by integrating sentiment analysis, topic modeling, and social network analysis. The goal is to reveal public sentiment trends, core concerns, and interaction patterns among users. First, we collected social media data related to the specified topic using web scraping tools, and performed sentiment classification using the NLPIR model, with a particular focus on the expression of negative sentiments. Next, we applied BERT-based tools for topic modeling on negative tweets, identifying key issues that the public is concerned about in discussions. Finally, through social network analysis, we explored the information dissemination paths among users and identified high-impact users who play critical roles in spreading negative sentiments. This study provides new insights into public opinion monitoring and crisis management on social media and proposes an effective data analysis framework, which can be further extended by incorporating multimodal analysis and cross-platform data in future research.