ABSTRACT
This document intends to analyze the sentiments underlying COVID-19 vaccination tweets. To achieve the objective, 38,034 publications from this social network are extracted through data mining, applying Machine Learning techniques, specifically sentiment analysis and network analysis, to identify the feelings expressed by Twitter users. We also identify the most relevant Twitter accounts on vaccination issues. The results suggest that feelings about vaccines are primarily negative;fear and anger, respectively, are the most recurring emotions in Twitter opinions. Moreover, we noted that the most relevant accounts belong to the media, politicians, and influencers, classified according to their feelings toward the vaccine. Opposition to the government with feelings of anger and opposition to recognized media with joyful emotions stand out. © 2022 Universidad de La Sabana. All rights reserved.
ABSTRACT
The effects of the different message strategies related to COVID-19 on the generation of eWOM were analyzed;that is, if the publications referring to the pandemic receive greater participation by users of social networks in Colombia. 562 company posts on Facebook were reviewed, of which 382 were subjected to the negative binomial regression model. It was found that no message strategy related to COVID-19 affects the rate of comments. The influence of different types of content on reactions and shared content was also identified. It is concluded that social networks are recreation and entertainment scenarios;therefore, the informative content does not generate impacts on the volume of comments, reactions, or share content.