Analysis of Online Public Opinion on Ecological Issues in Guangxi Based on the LDA Model

As ecological issues become increasingly severe, public attention and discussions on environmental protection topics have been widely spread through online platforms. This paper focuses on the online public opinion regarding water pollution issues in Guangxi, using video comments from Douyin and Bilibili as the data source. The LDA model is employed to extract the opinion topics, and sentiment analysis tools are used to explore the differences in public emotional tendencies. The study finds that there is little difference in the proportion of positive emotions in the comments on Douyin and Bilibili, while the proportion of negative emotions on Douyin is significantly higher than on Bilibili. Douyin comments tend to express emotions in a more sensational manner, while Bilibili comments place more emphasis on rational discussion and in-depth analysis. The study suggests that factors such as user age, platform comment mechanisms, information dissemination speed, and platform algorithms could be the possible causes of the differences in public opinion tendencies between the two platforms. The paper also discusses the impact of policy factors, media factors, and public participation on online public opinion and provides corresponding suggestions to guide opinion regulation bodies in managing online public sentiment.