This study aims at analysing sentiments and emotions expressed by Twitter users at the onset of the Covid-19 crisis. To this end, we collect a corpus of over 550,000 original tweets (13 million tokens), written in English and Italian. First, we analyse it through techniques of sentiment and emotion analysis. We find similarities, but also significant differences, between the English and the Italian corpora in the evolution of sentiment, and striking similarities in the distribution of emotions (trust and fear being the most pervasive in both languages). We then carry out a pilot qualitative analysis on two samples of data through the lens of Appraisal theory. This allows us to further interpret the results of the automated analysis, while also confirming the accuracy of our quantitative methods.
Sentiment and Emotion Analysis Meet Appraisal. A Corpus Study of Tweets Related to the COVID-19 pandemic
Claudia Roberta Combei
;
2021-01-01
Abstract
This study aims at analysing sentiments and emotions expressed by Twitter users at the onset of the Covid-19 crisis. To this end, we collect a corpus of over 550,000 original tweets (13 million tokens), written in English and Italian. First, we analyse it through techniques of sentiment and emotion analysis. We find similarities, but also significant differences, between the English and the Italian corpora in the evolution of sentiment, and striking similarities in the distribution of emotions (trust and fear being the most pervasive in both languages). We then carry out a pilot qualitative analysis on two samples of data through the lens of Appraisal theory. This allows us to further interpret the results of the automated analysis, while also confirming the accuracy of our quantitative methods.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
