Sentiment analysis of emotions in social media during the COVID-19 pandemic

In recent decades, social media has attracted increasing interest from researchers as a valuable source of data for analyzing users' opinions on specific topics. When extreme events occur, netizens' comments spread quickly online and can influence people's judgments both positively and negatively. Online social media such as Twitter provide a platform for netizens to learn about epidemics and express their opinions. Therefore, social media platforms can be a good tool for governments, businesses, and other organizations to learn about and respond to public concerns and emotional reactions with institutional messages or emotional comfort. During the Covid 19 pandemic, many companies used television commercials to build trust and hope in Italian families by emphasizing the principles of solidarity and fraternity and putting the brand in the background. In this paper, we attempt to measure the effectiveness of a change in the advertising approach of big brands using a specially created corpus of 20,982 tweets. Our results show that Internet users expressed an overall sense of belonging and gratitude during the lockdown, as well as an appreciation for television advertising that was perceived as disruptive at other times. We also became aware that the only currently available Italian lexicon for text analysis, developed by CNR, needs to be improved because it deals with formal Italian vocabulary and is not suitable for informal language such as tweets. The decoding of irony in a text by automatic polarity determination techniques remains a problem that needs further study.