SenWave: COVID-19 Tweets Labelled for Ten Sentiments Analyzed
SenWave: tweets about COVID-19 labeled for ten sentiments, by Qiang Yang and colleagues. The package holds 10,000 English tweets and 9,999 Arabic tweets that annotators labeled with any of ten emotions (optimistic, thankful, empathetic, pessimistic, anxious, sad, annoyed, denial, official report, joking), and Spanish, French and Italian versions of the 10,000 English tweets that the authors made with Google Translate, each carrying the English labels. The Arabic tweets also carry one of four overall sentiment labels (neutral 4,269, negative 2,750, positive 1,562, joking 1,418). That is 49,999 labeled tweets in five languages. The tweets were collected from 1 March 2020 (per the upstream README). The upstream repository also lists a large number of unlabeled tweet IDs; they are not part of these files. The files have no usable tweet IDs and no users, times or other fields.
Twitter (X)