In recent years,China’s online e-commerce platforms have continued to develop,and the rapid development of e-commerce sales platforms has brought new vitality to the field of agricultural product marketing.Among them,Tiktok Platform has carried out live broadcast activities of county heads to help farmers,connecting agriculture and the Internet platform,which has provided great impetus in promoting agricultural development,increasing agricultural product sales,increasing farmers’ income,narrowing the gap between urban and rural development,and has also continued to help achieve rural revitalization.Epidemic infectious diseases are frequent,under the new economic format,consumers are more dependent on online shopping platforms.The progress of network information technology has prompted consumers to form a lifestyle of online consumption and online shopping.The online review of target products is an important reference for consumers to make decisions,which provides new ideas for research in the field of agricultural product marketing.This paper takes the online comment data of Tiktok county magistrate’s live broadcast of agricultural assistance activities as the research object,which has important reference significance for the follow-up development of this activity.With the help of octopus software,this paper crawled the online comments of the county magistrate’s live broadcast of agricultural assistance activities on the Tiktok platform.After data cleaning,text segmentation and other pre-processing procedures are completed,the review text is first analyzed as a whole,and high-frequency words in the corpus are extracted to preliminarily understand the dimensions of consumers’ concerns;At the same time,the relationship between keywords is clarified through the social semantic network diagram.Secondly,the data is brought into the LDA clustering model to extract the subject words of the online comment text,and the ten major subject dimensions of the county magistrate’s live broadcast of agricultural assistance activities are summarized.Then,extract consumer satisfaction,quantify the review text data,compare the emotional analysis methods of SnowNLP and emotional dictionary,and select SnowNLP with higher accuracy to quantify the emotional value;The results are combined with LDA clustering results to obtain consumer comments;At the same time,the emotional value is converted into a five point system as a satisfaction index.Finally,based on SOR theory,a conceptual model of online comments on the county magistrate’s live broadcast agricultural activities on consumer satisfaction is built with grounded theory to clarify the specific impact mechanism of various dimensions on consumer satisfaction.The results are as follows:(1)Online review consumer attention factor are:express logistics,product quality,freshness,description compliance,merchant customer service,taste,packaging,repurchase intention,anchor characteristics and price perception.(2)There are 5532 comments with positive tendencies,accounting for more than 70%of the online comment data;There are 1968 comments with negative tendency in the data,accounting for 26.24%of the total comments.(3)Among the seven dimensions that affect consumer satisfaction,five dimensions,namely,value perception,attraction characteristics,quality,taste and taste,and logistics packaging,have a positive promoting effect;Business services and recommended buy backs have a negative regulatory effect,which is not conducive to the improvement of satisfaction.This paper proposes a set of frameworks from clustering themes and extracting emotions to establishing a conceptual model of consumer satisfaction,which provides a new reference for relying on online reviews to understand consumer satisfaction. |