Public Opinion Analysis of TikTok Hot Search Based on NLP

If the outburst of network public opinion is not controlled in time, it may even allow netizens to map the emotions vented online to the real society, causing economic and other losses to the real society. This paper aims to establish a public opinion analysis system based on NLP for Tiktok hot search, which can quickly grasp the online public opinion of Tiktok hot search, reflect the attitude of various regions in China towards the hot spot through data visualization, grasp the trend of public opinion, avoid the escalation of public opinion and prevent public opinion crisis. The NLP natural language processing algorithm is analyzed, and the implementation process of Douyin crawler and the sentiment analysis of comments are described in detail. In this system, the review information of Douyin hot search is obtained by Scrapy framework, data preprocessing is carried out, then the Jieba Chinese word segmentation technology is adopted to put the review data into the SnowNLP model for emotion analysis, and the sentiment value of the review is obtained, and the data is visualized, and then the Flask visualization screen is built.