Twitter discussions and emotions about COVID-19 pandemic: a machine learning approach

Identifying pandemic-related stress factors from social-media posts – Effects on students and young-adults

“…They showed that people were in a better mood after engaging with an OCSI than before. Another study conducted on Twitter discussions, identified popular unigrams, bigrams, salient topics and themes, and sentiments in collocated tweets (Xue et al, 2020). Another study about the impact of COVID-19 on students with disabilities or health concerns (Zhang et al, 2020) showed that people with disabilities and health concerns are more worried about classes going online than their peers without disabilities.…”

Understanding Public Sentiments, Opinions and Topics about COVID-19 using Twitter

“…Figure 3 shows the time series of the proportions of emotions exhibited by Twitter users in Singapore with regards to COVID-19. The week after the circuit breaker 1 started for Singapore on 7 April (April 9 to April 15), it can be seen that percentages of the emotions "fear" (11.9%) and "sadness" (9.3%) dropped to one of the lowest percentages for 1 Measures to reduce movement, activity and interactions in the country. their respective emotions throughout the whole duration of the dataset; while the percentages of the emotions "trust" (25.1%) and "joy" (15.7%) rose to one of the highest percentages for their respective emotions.…”

“…The COVID-19 outbreak is a global pandemic that has infected millions and claimed the lives of hundred thousands. Social media, e.g., Twitter, has proved to be valuable in times of such global pandemics, as it reveals the real-time sentiment and discussions of Twitter users [1]. Additionally, with the implementation of quarantine measures and the prohibition of social gatherings across the world, the usage of social media has soared to extraordinary levels.…”

Large-scale, Language-agnostic Discourse Classification of Tweets During COVID-19

“…n-gram statistics) content analysis of Twitter chatter for gaining relevant insights [21,23,24,25,26,27,28,29], while other studies utilize computational approaches such as topic modeling [19,20,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49]. A high percentage of studies performing topic modeling and topic discovery on Twitter utilize the well-established Latent Dirichlet Allocation (LDA) algorithm [20,30,33,34,36,37,40,41,42,43,44,45,46,49]. Similar unsupervised approaches of word/n-gram clustering [38,39,47] or clustering of character/word embeddings [35,…”

“…Tweet data utilized for most of these studies are restricted to a single language. Majority of the studies restrict their analysis only to English tweets [19,24,29,30,33,37,39,41,43,44,45,46,48], possibly exacerbating the already existing selection bias. Other studies have restricted their datasets to Japanese [47], Korean [21], Persian/Farsi [36], and Polish [31] tweets.…”