What social media told about us in the time of COVID-19: a scoping review

Shu-Feng Tsao, MSc1, Helen Chen, PhD1, Therese Tisseverasinghe, MLIS2, Yang Yang,
MSc1, Lianghua Li, BSc3, Zahid A. Butt, PhD1
1School of Public Health and Health Systems, University of Waterloo, Waterloo, Ontario,
Canada
2Seneca College, King City, Ontario, Canada
3Faculty of Science, University of Waterloo, Waterloo, Ontario, Canada

Correspondence to:

Zahid Ahmad Butt

School of Public Health and Health Systems
University of Waterloo,
200 University Avenue West,
Waterloo, ON, N2L 3G1
Tel: 519-888-4567 ext. 45107
E-mail: zahid.butt@uwaterloo.ca
Summary

With the onset of COVID-19 pandemic, social media has rapidly become a crucial
communication tool for information generation, dissemination, and consumption. In this
scoping review, we selected and examined peer-reviewed empirical studies relating to
COVID-19 and social media during the first outbreak starting in November 2019 until May
2020. From an analysis of 81 studies, we identified five overarching public health themes
concerning the role of online social platforms and COVID-19. These themes focused on: (i)
surveying public attitudes, (ii) identifying infodemics, (iii) assessing mental health, (iv)
detecting or predicting COVID-19 cases, (v) analyzing government responses to the
pandemic, and (vi) evaluating quality of health information in prevention education videos.
Furthermore, our review highlights the paucity of studies on the application of machine
learning on social media data related to COVID-19 and a lack of studies documenting real-
time surveillance developed with social media data on COVID-19. For COVID-19, social
media can play a crucial role in disseminating health information as well as tackling
infodemics and misinformation.

Introduction

The Corona Virus Disease 2019 (COVID-19), or Severe Acute Respiratory Syndrome Novel
Coronavirus (SARS-COV-2), is a significant international public health issue. As of
November 14, 2020, an estimated 53.5 million people around the world have been infected
with the virus, with about 1.3 million deaths. 1 As a consequence of the pandemic, social
media is becoming the platform of choice for public opinions, perceptions, and attitudes
towards various events or public health policies regarding COVID-19. 2 Social media has
become an even more pivotal communication tool for governments, organizations, and
universities to disseminate crucial information to the public. Numerous studies have already
used social media data to help identify and detect infectious disease outbreaks and to
interpret public attitudes, behaviours, and perceptions.3-6 Social media, particularly Twitter,
may be used to explore multiple facets of public health research. A recent systematic review
identified six categories of Twitter use for health research, namely, content analysis,
surveillance, engagement, recruitment, interventions, and frequency analysis of Twitter
users.4 However, this review only included broader research terms such as health, medicine,
or disease using Twitter data and did not focus on specific disease topics such as COVID-
19. Another article analyzed tweets on COVID-19 and identified 12 topics categorized into
four main themes: the origin of COVID-19, the source of the coronavirus, impact of COVID-
19 on individuals, and countries, and methods of decreasing the spread of COVID-19. 7 In
this study data were not available on COVID-19 related tweets before February 2020 thereby
missing the initial part of the epidemic, and the data on tweets were limited to Twitter only
within a one-month period.

Social media can also be effectively used to communicate health information to the general
public during a pandemic. Emerging infectious diseases (EID), such as COVID-19, almost
always results in increased usage and consumption of media of all forms by the general
public for information.8 Therefore, social media plays a vital role in people’s perception of
disease exposure, resultant decision making and risk behaviours.9,10 As information on
social media is generated by users, such information can be subjective or inaccurate, and
frequently includes rumours, misinformation, and conspiracy theories. 11 Hence, it is
imperative that accurate and timely information is disseminated to the general public about
emerging threats such as the novel coronavirus. A systematic review explored the major
approaches used in published research on social media and EID.12 The review identified
three major approaches: (a) assessment of the public’s interest in and responses to EID, (b)
examination of organizations’ use of social media in communicating EID, and (c) evaluation
of the accuracy of EID related medical information on social media. However, this review
did not focus on studies that used social media data to track and predict EID outbreaks.

Analyzing and disseminating information from peer-reviewed published research can guide
policymakers and public health agencies to design interventions for accurate and timely
knowledge translation to the general public. Therefore, keeping in view the gaps mentioned
above, we conducted a scoping review with the aim of understanding what roles social media
have been playing since the beginning of the COVID-19 crisis. We wanted to know
specifically about public attitudes and perceptions towards COVID-19 on social media,
information about COVID-19 on social media, social media use for prediction and detection
of COVID-19, the impact of COVID-19 on mental health and government responses to
COVID-19 on social media. Our objective was to identify studies on social media related to
COVID-19 that focused on the following: (1) Infodemics (2) Public attitudes, (3) Mental
health, (4) Detection or prediction of COVID-19 cases, (5) Government responses to the
pandemic, and (6) Quality of health information in videos.
Methods

Overview

Studies exploring the use of social media on COVID-19 were reviewed using the scoping
review methodology of Arksey and O’Malley13 and Levac et al.14 The five-step scoping
review protocol was followed, and the Preferred Reporting Items for Systematic Reviews
and Meta-Analyses (PRISMA) extension for scoping reviews was employed in this study.

Data Sources

Exploratory searches were conducted on COVID-19 Open Research Dataset Challenge

(CORD-19) as well as Google Scholar in April 2020. These searches helped define the
review scope, develop the research questions, and determine eligibility criteria. After such
activity, MEDLINE/PubMed, Scopus, and PsycINFO were selected for this review since
they include peer-reviewed literature in the fields of medicine, behavioural sciences,
psychology, healthcare system, and clinical sciences. Since the start of the current pandemic,
COVID-19 articles were reviewed and published at an unprecedently rapid rate with
numerous publications that were peer-reviewed and available ahead of print, referred to as
"pre-prints" or "articles in press." In this review, we consider peer-reviewed pre-prints
equivalent to normal peer-reviewed articles, and relevant articles were screened accordingly.

Search Strategy and Selection Criteria

The search strategies utilized index terms, where applicable, and free-text terms to capture
the following two concepts: (a) social media, including both general terms and specific
platform names (e.g., Twitter, Facebook, Weibo, YouTube); (b) COVID-19. For each
database, both indexed terms (i.e., MeSH, Emtree) and natural language keywords were used
with Boolean operators (i.e., AND, OR, NOT) and truncations. Table 1 shows keywords,
indexed terms, and their variations used for the literature search. Since each database has
distinctive search functionality, individually tailored search statements were developed with
appropriate search filters for each database. Final search statements, along with a list of
search results, were downloaded from each database.
 Table 1: Keywords, indexed terms, and their variations used in the literature search

Keywords Variations
betacoronavirus; severe acute respiratory syndrome; covid 2019; corona-19; n-cov; novel
COVID-19
coronavirus; sars-cov; wuhan 2019
twitter; tweet; retweet; facebook; weibo; sina; youtube; webcast; user comment; online
social media
post; online discussion; social network; social media; online community; mobile app
Indexed terms Variations
betacoronavirus; coronavirus infections; severe acute respiratory syndrome; coronavirus
COVID-19
disease 2019; covid-19
social networking; social media; mobile applications; blogging; social networking (online);
social media
online social network; webcast; mobile applications; mobile computing; social network

Articles were included if (a) they discussed the use of social media for COVID-19 research,
and (b) were original, empirical studies. Only peer-reviewed articles, including peer-reviewed
preprints, in English or Chinese, were included. A decision to include Chinese publications
was based on the fact that COVID-19 cases were first reported in Wuhan, China and many
initial and relevant studies were published in Chinese; therefore, we wanted to capture a
majority of studies that were on the use of social media for COVID-19 research. All articles
published between November 1, 2019, and November 4, 2020, were included. Publications
such as reviews, opinion pieces, pre-prints, books, book chapters and any empirical study that
were not peer-reviewed or written in languages other than English or Chinese were
automatically excluded. The final reference list was generated based on originality and
relevance to the broad scope of this review.

Screening Procedure

Primary reviewer (SFT) mainly carried out title and abstract screening for each article to
determine if an article met the inclusion criteria. If the criteria were confirmed, then it was
included; otherwise, it was excluded. Articles were broken into paragraphs to contain one code
in each section. Next, quotes were sorted under each code, applying Ose’s method. 15 Braun
and Clark’s thematic analysis method was employed and involved searching for the text that
matched the identified predictors (codes) in from the quantitative analysis and discovering
emergent codes that were relevant to either the study objective or identified in the relevant
literature review.16 Finally, we categorized the codes into main themes. These codes and
themes were compared and clarified to draw conclusions around the main themes by SFT, ZB,
and RY. SFT is fluent in both English and Chinese; the secondary reviewer (ZB) is fluent in
English, and the tertiary reviewer and domain expert (YY, HC) are both fluent in English and
Chinese. Any discrepancies among reviewers were discussed with the research team to reach
consensus.

Results

Figure 1 illustrates the PRISMA flow diagram of the literature search. With the application of
appropriate search filters, a total of 2405 articles were retrieved from the identified databases:
PubMed (n=1084), Scopus (n=1021) and PsycINFO (n=300). Among these, 670 duplicates
were excluded. Of the remaining 1735 articles, 1434 were deemed non-empirical, such as
comments, editorial essays, letters, opinions, and reviews. This left 301 articles for a full-text
review based on the screening results of titles and abstracts. After the full-text review, 81
articles were included in this scoping review.
Identification

2405 articles identified from database

search
670 duplicates
removed
Screening

1735 articles eligible for title and

abstract screening
1434 articles
excluded
Eligibility

301 articles qualified for full-text

review
220 articles excluded
Inclusion

81 articles selected

Figure 1: PRISMA flow chart of article extraction from the literature search
Table 21 summarizes the 81 selected articles on COVID-19 and social media. All articles were
written in English. Twitter (n=45) and Weibo (n=16) undoubtedly were the most studied social
media data. To categorize these chosen articles, we adopted a novel framework called “Social
media and Public Health Epidemic and REsponses (SPHERE) and developed a modified
version of SPHERE framework to organize the themes for our scoping review, as shown in
Figure 2.17 Themes were identified through reviewers’ consensus based on our modified
SPHERE framework. We identified six themes: (i) infodemics, (ii) public attitudes, (iii) mental
health, (iv) detecting or predicting COVID-19 cases, (v) government responses, and (vi) quality
of health information in videos.

Figure 2: Modified SPHERE Framework

Social Media as Contagion and Vector

1. Infodemics

According to the World Health Organization (WHO), the term “infodemic,” a combination of
“information” and “epidemic,” refers to a fast and widespread of both accurate and inaccurate
information about something, such as a disease like COVID-19.18 Twelve articles tackled
COVID-19 related infodemics circulating on social media platforms. Rovetta and
Bhagavathula analyzed over 2 million queries from Google Trends and Instagram between

1
Table 2 is displayed at the end of this paper
February 20 and May 6, 2020. Their findings revealed that as global interest of COVID-19
information increased, so did its infodemic.19 Gallotti et al analyzed over 100 million tweets
and found that even before the onset of the COVID-19 pandemic, entire Inlay’s expose
measurable waves of infodemic has threatened public health.20 Pulido et al sampled and
analyzed 942 tweets which revealed that while false information had higher number of tweets,
it also had less retweets and lower engagement than tweets comprising scientific evidence or
factual statements.21 Kouzy et al22 set out to investigate the extent to which misinformation or
unverifiable information about the COVID-19 pandemic was spread on Twitter by analyzing
673 English tweets. Their results showed that misinformation accounted for 24.8% of all
tweets. Healthcare or public health accounts had the lowest rate; yet still 12.3% of their tweets
included unverifiable information. Moscadelli et al collected and reviewed 2,102 news articles
circulated on the internet. Their analysis revealed that fake news accounted for 23.1% of
articles studied which were shared over 2 million times between December 31, 2019 and April
30, 2020.23 Similarly, another quantitative study by Galhardi et al comparing the proportion of
fake news shared on WhatsApp, Instagram, and Facebook in Brazil revealed that fake news
was predominately shared on WhatsApp.24 A UK study by Ahmed et al, analyzed 22,785 tweets
posted by 11,333 Twitter users with “#FilmYourHospital” hashtag to identify and evaluate the
source of conspiracy theory on Twitter. Their work uncovered that ordinary people were the
major driver behind the spread of conspiracy theories.25 Another research investigated the 5G
and COVID-19 conspiracy theory circulating on Twitter with a random sample of 233 tweets.
The content analysis showed that 34.8% tweets linked 5G and COVID-19 and 32.2%
condemned such theory 26. A similar research by Bruns et al investigated 89,664 distinct
Facebook posts in Australia related to such conspiracy from 1 January to 12 April 2020 using
time series and network analysis.27 The results showed that this conspiracy went viral after
March 19, 2020 with unusual coalition among various groups on Facebook. Islam et al
analyzed 2,311 COVID-19 related infodemic reports from December 31, 2019 to April 5, 2020
and demonstrated misinformation was mainly driven by rumors, stigma, and conspiracy
theories circulating on various social media and other online platforms.28 Associations between
infodemic and bot activities on social media is another important research direction. One study
analyzed 12 million tweets from the US and 15 million tweets from the Philippines from March
5 to 19, 2020 and both countries showed a positive relation between bot activities and rate of
hate speech in denser and more isolated communities.29 Brennen et al qualitatively analyzed
96 samples of visuals from January through March 2020 and categorized misinformation into
6 trends, and found, fortunately, there have been no involvement of AI “deepfakes” techniques
so far.30

Social Media for Surveillance and Monitoring

Three themes emerged under this category: public attitudes, mental health, and detection or
prediction of COVID-19 cases. The former two themes are reflections regarding the public
perceptions and mental health impacts of the pandemic; the latter includes typical surveillance
studies aiming to propose ways for the detection or prediction of the COVID-19 cases.

1. Public attitudes

Forty-eight selected articles gauged the attitudes and emotions expressed by social media users
regarding the COVID-19 pandemic mainly by employing content and sentiment analysis.
Twitter and Weibo accounted for 34 and 8 articles, respectively. This theme can be further
divided into the following sub-themes: (1) Public sentiment toward the COVID-19 pandemic
and interventions, (2) stigma and racism, and (3) ageism.

1.1 Public sentiment

To learn what the public’s perspectives about the overall COVID-19 pandemic and its
interventions, Alaa Abd-Alrazaq et al.7 analyzed 167,073 unique English tweets divided into
four categories: origin of the COVID-19 virus; sources of the virus; its regional and global
impact on people and society; and methods to reduce its infection. Tweets regarding economic
loss had the highest average number of likes whereas travel bans and warning had the lowest.7
Kwon et al investigated 259,529 English tweets in the USA using trending and spatiotemporal
analyses and uncovered tweets about social disruptiveness had highest number of retweets and
tweets about COVID-19 interventions had highest number of likes.31 A content analysis of 522
Reddit comments showed that “Symptoms” accounted for 27% of all comments, followed by
“Prevention” (25%).32 Likewise, another content analysis of 155,353 unique English tweets
show the most mentioned topic was “peril of COVID-19.”33 Still another research examined
126,049 English tweets using sentiment analysis and latent Dirichlet analysis (LDA) for topic
modeling revealed the most common emotion mentioned was fear, and the most common topic
mentioned was the economic and political effects.34 Al-Rawi et al studied emojis in over 50
million tweets and identified 5 primary subjects: “morbidity fears,” “health concerns,”
“employment and financial issues,” “praise for frontline workers,” and “unique gendered emoji
use.”35 Samuel et al investigated 293,597 tweets with sentiment analysis and found more
positive emotions toward the US economy re-opening than negative emotions. 36 Analyzing
2,558,474 English tweets using clustering and network analyses, Odlum et al found African
Americans shared positive sentiments and encouraged virtual discussions and prevention
behaviours.37 A study investigated gender differences in terms of topics by analyzing 3,038,026
English tweets.38 The results demonstrated that tweets from females were more likely about
family, social distancing, and healthcare, whereas tweets from males were more about sports
cancellations, pandemic severity, and politics. In Canada, Xue et al analyzed 1,015,874 tweets
via LDA to identify nine themes about family violence.39 In Australia, Yigitcanlar et al
analyzed 96,666 tweets and found that the public’s attitude could be captured efficiently. 40 One
qualitative content analysis of 30 profiles from Instagram, Twitter, and TikTok in Brunei
resulted in five types of attitudes toward social distancing: fear, responsibility, annoyance, fun,
and resistance.41 Arpaci et al in Turkey investigated 43 million tweets from March 22 to 30,
2020 via evolutionary clustering analysis. Their findings showed that words like “death”,
“test”, “spread”, and “lockdown” appeared more frequently, suggesting people are afraid of
infection, and death, and staying at home due to such fear, and supportive of social
distancing.42 A study conducted in Luzon, Philippines,43 where sentiment analysis was
employed using natural language processing (NLP), revealed that most Filipino Twitter users
expressed negative emotions towards COVID-19, and the negative mood grew stronger
overtime under the lockdown.43 Sentiment analysis of 107,990 English tweets uncovered that
negative feeling toward COVID-19 pandemic dominated in Thailand, and topic modeling
revealed three major themes in people’s concerns: “the COVID-19 pandemic emergency,”
“how to control COVID-19,” and “reports on COVID-19.”44 Another research analyzed
373,908 tweets and retweets in Belgian which showed that the public focused more on EU
coalition to tackle the pandemic.45 De Santis et al analyzed 1,044,645 tweets to identify daily
hot topics related to COVID-19 pandemic in Italy and developed a framework for prospective
research.46 One thematic analysis study of 1,920,593 Arabic tweets in Egypt revealed that
negative emotions and sadness were loaded in tweets showing affective discussions, and the
dominant themes included “the outbreak of the pandemic,” “metaphysics responses,” “signs
and symptoms in confirmed cases,” and “conspiracism.”47 In Singapore, Lwin et al examined
20,325,929 tweets using sentiment analysis and showed the public emotions shifted over time:
from fear to anger and from sadness to grateful.48 Chang et al examined over 1.07 million
Chinese texts from various online sources in Taiwan using deductive analysis and found
negative sentiments mainly came from online news with stigmatizing languages linked with
the COVID-19 pandemic.49 In India, one study investigated 410,643 tweets via sentiment
analysis and LDA and showed that positive emotions were overall significantly higher than
negative sentiments, but this observation diminished at individual levels. 50 Another research
analyzed 29,554 tweets from lockdown 2.0 and 47,672 tweets from lockdown 3.0 via sentiment
analysis uncovered positive attitudes toward lockdown 2.0 but turned to negative attitudes
toward lockdown 3.0 in India.51 One study analyzed 868 posts from Reddit and found
sentiments to be 50% neutral, 22% positive, and 28% negative in India.52 A study in South
Korea examined 43,832 unique users and their relations on Twitter using content and network
analyses demonstrated tweets including medical news were more popular.53 A research from
Ireland analyzed 203,756 tweets through topic modeling and identified that “WAR” was the
most common used frame.54 In the USA, Damiano et al qualitatively analyzed 600 English
tweets and showed neutral sentiment. Politics also played an essential role in shaping people’s
opinion.55 A research studied 19,803 tweets from Democrats and 11,084 tweets from
Republicans using RF in the USA and showed that Democrats emphasized more on public
health threats to American workers, whereas Republicans emphasized more threats from China
and business.56 Results of a study involved various online data sources from Italy, UK, USA,
and Canada showed media was the major driver of the public’s attention, despite rapid
saturation.57 Compared with other users, Reddit users focused more on disease data and
prevention interventions. Researchers in Spain studied 22,223 tweets using topic modeling and
network analysis.58 They identified eight frames and found that the entire pandemic could be
divided into three periods: “pre-crisis,” “lockdown” and “recovery” periods. On the other hand.
Jelodar et al proposed a novel method to detect meaningful latent-topics and sentiment-
comment-classification on COVID-19–related posts using 563,079 COVID-19–related
English Reddit posts.59 Samuel et al examined over 900,000 tweets to study the accuracy of
tweet classifications among logistic regression and Naïve Bayes (NB) methods. They found
that NB had 91% of accuracy compared with 74% from the logistic regression model.60

Another study investigated how the Chinese government used the popular social media
platform, Sina Weibo, to promote citizen engagement during the COVID-19 crisis.61 Xuehua
Han et al.62 analyzed 1,413,297 Weibo posts and found that the public reacted sensitively to
the epidemic, especially in metro areas.62 Zhao et al studied 4,056 topics from the Sina
Microblog hot search list and found the public emotions shifted from negative to neural to
positive over time, and five major public concerns63. Jiawei Li et al.64 conducted an
observational infoveillance study with a linear regression model by analyzing 115,299 Weibo
posts. The results showed that the number of Weibo posts positively correlated with the number
of reported cases of coronavirus in Wuhan. In addition, the qualitative analysis classified the
topics into the following four overarching themes: (a) cause of the virus; (b) epidemiological
characteristics of the COVID-19; (c) public responses; and (d) others. 65 Qiang Chen et al.
examined relationships between citizen engagement through government social media
(CEGSM) and media richness, dialogic loop, content type, and emotion valence.61 GEGSM
refers to sum of shares, likes, and comments in this study, so the higher the sum, the greater
the GEGSM. Media richness quantifies how much a sender transfers information to a receiver
via a medium based on the media richness theory. Dialogic loop, or dialogic communication
theory, is defined as an approach that promotes a dialogue between a speaker and audience.
According to the American Psychological Association (APA), emotion valence refers to “the
value associated with a stimulus expressed on a continuum from pleasant to unpleasant or from
attractive to aversive”.65 For instance, happiness is typically considered pleasant valence. The
researchers analyzed 1,411 COVID-19 related posts from "Healthy China," an official account
of the National Health Commission of China (NHCC) on Weibo. Findings showed an inverse
association between media richness and CEGSM, indicating that posts with plain texts had
higher CEGSM than posts with pictures or videos. A positive association between dialogic
loop and CEGSM was found, as evidenced by 96% posts using hashtags and 25% employing
questions. CEGSM also had a positive relationship with emotion valence, as evidenced by
higher CEGSM with positive emotions shown in posts.61 The last chosen article proposed a
new multiple-information susceptible-discussing-immune (M-SDL) model to analyze the
public opinion propagation of the COVID-19 from Weibo posts collected from December 31,
2019, to February 27, 2020. The researchers reported that the reproduction rate of this proposed
M-SDL model reached 1.78 in the early stage of COVID-19 but decreased to and maintained
around 0.97. Such a result demonstrated that the information on COVID-19 would continue to
erupt slowly in the future until it stabilized without violent information outbreak.Wang et al
analyzed 999,978 randomly selected COVID-19 related Weibo posts through unsupervised
BERT model for sentiments and TF-IDF model for topic modeling. They identified four public
concerns: “the virus origin,” “symptom,” “production activity,” and “public health control” in
China.66 Xi et al examined 241 topics with their views and comments via thematic and temporal
analysis and found the “Contributing to the community” theme was the most dominant in the
first phase (January 20–February 20, 2020).67 The “Older patients in hospitals” theme was most
prevailing in the second (February 21–March 17, 2020) and third phase (March 18–April 28,
2020). Employing Wilcoxon tests, Su et al examined posts from 850 Weibo users and 14,269
Tweets from Italy.68 The findings showed Italians paid more attention to leisure whereas
Chinese paid more attention on group, religion, and emotions after lockdowns. Analyzing top
200 accounts from WeChat via regressions and content analysis, Ma et al found positive
impacts of nonmedical and medical reports on people’s behaviours. 69 Employing Kendall’s
Tau-B rank test, Xie et al investigated relations among Baidu Attention Index (BAI), daily
Google Trends, and numbers of COVID-19 cases and deaths. 70 Daily GT was correlated to 7
indicators, whereas daily BAI was only correlated to 3 indicators. 71 Eventually, Zhu et al
analyzed 1,858,288 Weibo posts and found topics changed over time, but political and
economic posts attracted greater attentions.71

1.2 Stigma and Racism

Bumsoo Kim72 analyzed 27,849 individual tweets in South Korea using a binary logistic
regression to gauge network size, and semantic network analysis to capture contextual and
subjective factors. The results indicated that social network size was inversely correlated with
impolite language use. Namely, users with larger social networks were less likely to post
uncivil messages on Twitter. This study suggested that the social network size influenced social
media users' language choice in their postings.72 A research compared pre- and post-
introduction of “Chinese virus” or “China virus” influenced the public stigma with 16,535
English tweets from pre-introduction period and 177,327 tweets from post-introduction
period.73 The results showed an almost 10 times increase nationwide and statewide from 0.38
tweets referencing those two terms posted per 10,000 people in the pre-period to 4.08 tweets
post-period in the USA. Another similar study examined 339,063 tweets from non-Asian
respondents via local polynomial regression and interrupted time-series analysis.74 The
findings demonstrated that when stigmatizing terms like “Chinese virus” were used by media
starting from March 8, 2020, the bias index—Implicit Americanness Bias—began to increase,
and such bias was more profound in conservatives. Nguyen et al analyzed 3,377,295 race-
related tweets in the USA using sentiment analysis uncovered a 68.4%. increase in negative
tweets referring Asians whereas tweets referring other races remained stable.75

1.3 Ageism
A study76 investigating Twitter content related to both COVID-19 and older adults analyzed a
random sample of 351 English tweets. It turned out that 21.1% of the tweets implied diminished
regard for older adults by downplaying or dismissing concerns over the high fatality of
COVID-19 on this population.76 Another similar research examined 188 tweets via thematic
analysis and revealed 90% against ageism, whereas 5% back up and 5% neutral.77

2. Mental Health

Two of the 81 reviewed studies, both based in China, focused on assessing social media users’
mental health.78,79 A cross-sectional study78 investigated the relationship between anxiety and
social media exposure (SME), theoretically defined as “the extent to which audience members
have encountered specific messages”.80 The researchers distributed an online survey based on
the Chinese version of WHO-Five Well-Being Index (WHO-5) for depression and the Chinese
version of generalized anxiety disorder scale (GAD-7) for anxiety. Respondents included 4,872
Chinese citizens aged 18 and over from 31 provinces and autonomous regions in China. After
controlling for all covariates through a multivariable logistic regression, the finding uncovered
that frequent SME increased the odds ratio of anxiety, demonstrating frequent SME potentially
contributing to the mental health problems during the COVID-19 outbreak.78 To explore how
people's mental health was influenced by the COVID-19, Li et al.798 analyzed posts from
17,865 active Weibo users to compare pre- and post-declaration of COVID-19 on 20 January,
2020. The researchers uncovered increased negative sentiments including anxiety, depression,
and indignation post-declaration, and decreased positive sentiments expressed in Oxford
happiness score. Additionally, cognitive indicators increased in social risk judgment but
decreased in life satisfaction post-declaration.78

3. Detection or prediction of COVID-19 cases

Six of the 81 studies investigated the prediction of COVID-19 outbreaks using social media
data. Qin et al.79 attempted to predict the number of new suspected or confirmed COVID-19
cases by collecting social media search indexes (SMSI) for symptoms such as dry cough, fever,
chest distress, coronavirus, and pneumonia. The data were analyzed using subset selection,
forward selection, lasso regression, ridge regression, and elastic net. Results showed the
optimal model was constructed via the subset selection. The lagged SMSI was a predictor of
new suspected COVID-19 cases and could be detected 6-9 days before the new case
confirmation77. To evaluate the possibility of early prediction of the COVID-19 cases via
internet searches and social media data, Li et al.82 used keywords “coronavirus” and
“pneumonia” to retrieve corresponding trend data from Google Trends, Baidu Index, and Sina
Weibo Index. Using the lag correlation, the results showed that for new lab-confirmed cases,
the highest correlation was found 9-12 days earlier for the keyword "coronavirus" in the three
platforms. Similarly, for the new suspected cases, the highest correlation was found 6-8 days
earlier for “coronavirus.” The keyword “pneumonia” for new suspected cases had the highest
correlation eight days earlier across the three platforms. 82Peng et al studied 1,200 Weibo
records using spatiotemporal analysis, kernel density analysis, and ordinary least square
regression and found that “scattered infection”, “community spread”, and “full-scale outbreak”
were three phases of early COVID-19 transmission in Wuhan, China. 83 Elderly accounted for
over 50% of help seeking on Weibo, thus of high risk. To identify COVID-19 patients with
poor prognosis, Liu et al, analyzed Weibo messages from 599 patients along with telephone
follow-ups.84 The findings suggested risk factors involving older age, diffuse distribution, and
hypoxemia. A regression study analyzed GT searches, Wikipedia page views, and tweets and showed
that using current Wikipedia page views, one‐week‐ahead tweets, and two‐week ahead GT searches
were included to model the number of COVID-19 cases.- To model the number of deaths, all three
variables should be one-week further ahead85.

Social Media as Disease Control

1. Government Responses

To make the public inoculated against misinformation, public health organizations should
create and spread accurate information on social media because social media have increasingly
played an important role in policy announcement and health education. Six of the 81 articles
were categorized as “government responses” because they examined how government
messages and health education material were generated and consumed on social media
platforms. Two studies analyzed data from Sina Weibo,86,88 and the other four studies analyzed
data from Twitter.88,89

Zhu et al.86 conducted a study to measure the attention of Chinese netizens—“citizen of the
net” according to Michael F. Hauben90—to COVID-19 by analyzing 1,101 Weibo posts. They
found that Chinese netizens paid little attention to the disease until the Chinese government
acknowledged and declared the COVID-19 outbreak on January 20, 2020. Since then, high
levels of social media traffic occurred when Wuhan began its quarantine (January 23-24, 2020),
during a Red Cross Society of China scandal (February 1, 2020), and following the death of
Dr. Li Wenliang (February 6-7, 2020).86 Lifang Li87 collected 36,746 Weibo posts to identify
and categorize the situational information using support vector machines (SVM), naive Bayes
(NB), and RF, as well as features in predicting the number of reports using linear regression.
Except for posts categorized as “counter-rumours,” they found the more word counts, the more
reposts for all the other types. Likewise, posts from unverified users had more reposts in all
types, excluding the counter-rumours. For the counter-rumours, reposts would increase if its
users have a greater number of followers and from urban areas.87 A qualitative content analysis
was conducted to investigate how G7 leaders used Twitter concerning the COVID-19
pandemic by collecting 203 tweets.88 The findings showed that 82.8% were “informative,”
28.6% of them linked to official government resources; 9.4% were "morale-boosting;” and
6.9% were “political”.88 To assess the political partisan polarization in Canada regarding
COVID-19, Merkley et al.89 random sampled 1,260 tweets from 292 federal members of
Parliament’s (MP) social media and collected 87 Google search trends for the search term
"coronavirus". A survey was also conducted with 2,499 Canadian respondents aged 18 years
and above. The results demonstrated that regardless of party affiliation, MP members
emphasized the importance of social distancing measures and proper hand hygiene practices to
cope with the COVID-19 pandemic without tweets exaggerating concerns or misinformation
about COVID-19. Search interest in coronavirus among municipalities was strongly
determined by socioeconomic and urban factors rather than Conservative party vote share.89
Sutton et al studied 149,335 tweets from public health, emergency management and elected
officials and found that messages changed positively and negatively over time.91 Besides,
Wang et al investigated 13,598 COVID-19 related tweets via temporal and network analyses. 92
They categorized 16 types of messages and identified inconsistent and incongruent messages
expressed in 4 critical prevention topics: mask wearing, risk assessments, stay at home order,
and disinfectants/sanitizers.

2. Prevention Education in Videos

Eight chosen studies investigated the quality of Youtube videos with COVID-19 prevention
information. Basch et al.93 conducted a cross-sectional study and retrieved the top 100 YouTube
videos with most views uploaded in January 2020 with keywords "Coronavirus" in English,
with English subtitles, or in Spanish. These 100 videos together generated over 125 million
views. However, less than 33 videos included any of seven key prevention behaviours
recommended by the US CDC.93 The same group of researchers conducted a follow-up study
with the same criteria and a successive sampling design to gather top 100 mostly viewed
YouTube videos in January and in March 2020, respectively. Findings showed that the January
sample generated over 125 million views, and the March sample had over 355 million views.
Yet less than 50- videos in either sample contained any of the prevention behaviours
recommended by the US CDC.94 Additionally, a research investigated the first 100 “DIY hand
sanitizer” videos on YouTube with the most views and showed the average number of daily
pediatric poisoning calls increased significantly in March 2020 compared with the previous 2
years.95

To analyze the information quality on YouTube about the COVID-19 pandemic and to compare
the contents in English and Chinese Mandarin videos, Priyanka Khatri et al.39 collected a total
of 150 videos with the keywords “2019 novel coronavirus,” “Wuhan virus” in English and in
Mandarin. The DISCERN score and the medical information and content index (MICI) were
calculated as a reliable way to measure the quality of health information. It turned out the
average DISCERN score for reliability was low: 3.12 out of 5 for English videos and 3.25 for
Mandarin videos. The average cumulative MICI score of useful videos was also undesirable:
6.71 out of 25 for English videos and 6.28 for Mandarin videos. 96 In Spain, a similar research
studied 129 videos in Spanish and found such videos were usually incomplete and differed
according to different uploaders. Likewise, one study in South Korea found misleading videos
accounted for 37% of all collected 105 most viewed, and such videos had more likes, fewer
comments, and longer viewing times than useful videos. Two studies in Turkey investigated
quality of YouTube videos regarding COVID-19 information in dentistry. 99,99 One analyzed
top 116 English videos with at least 300 views and showed moderate quality and useful
information from these videos.99 The other one, however, uncovered poor quality for 24 out of
55 (43.6%) English videos, whereas good quality accounted for only 2 (3.6%) videos. 100

Discussion

Studies on social media data revealed our attitudes and mental state to a certain extent during
the COVID-19 crisis. These studies also showed how we generated, consumed and propagated
information on social media platforms when facing the rapid spread of the novel coronavirus
and extraordinary measures for the containment. In our review, public attitudes accounted for
nearly 60% of the reviewed articles. In terms of social media platforms, 46% of the chosen
articles used data from Twitter, followed by Weibo (18%). Machine learning analyses, such as
LDA and RF, were applied in research that studied public attitudes.

We identified six themes based on our modified SPHERE framework, including (i) infodemics,
(ii) public attitudes (iii) mental health, (iv) detection or prediction of COVID-19 cases, (v)
government responses to the pandemic, and (vi) quality of prevention education videos.
However, a common limitation in all chosen studies using social media data is the “data
quality” challenge, such as various formats, metrics, or even the definition of a variable. For
instance, the definition of a “view” on one social media platform is likely different from
another. Besides, not every social media offers accessible data like Twitter and Weibo. To
address these challenges, the selected studies have controlled for many factors, including social
media platforms, languages, locations, time, misspellings, keywords, or hashtags. However,
such search strategies resulted in many study limitations, like non-representative sample sizes
or selection bias, cross-sectional or retrospective study design. We also observed that, given
the large amount of available data, most studies across all domains still sampled smaller data
size for analyses, except four studies under the “public attitudes” theme analyzing over one
million posts via machine learning methods. Additionally, data from Twitter and Weibo
accounted for over 60% of our selected studies. Research examining other social media
platforms, including Facebook, Instagram, TikTok, Snapchat, and WhatsApp, are limited due
to barriers of data availability and accessibility. We also identified future research topics
needed for each category as we are currently in the midst of a second wave of the COVID-19
pandemic . From “infodemics” perspective, additional research is needed to investigate how
effective misinformation/rumours/fake news, such as anti-mask wearing, undermine
preventions to compromise public health, although social media companies like Twitter and
Facebook recently started to remove misinformation-based accounts or label unaccountable
posts. Robot (or bot) posts is another focus to be addressed. Studies evaluating effective
counter-infodemic interventions are also needed.

Articles regarding “public attitudes” toward the COVID-19 pandemic have revealed sentiments
that shifted over time. Yet, this theme can be a useful indicator when evaluating interventions,
social distancing and wearing masks for instance, that aim to reduce the COVID-19 infection
risks. However, public sentiments have not been incorporated into many intervention studies
by the time we conducted this review. When a disease like COVID-19 starts spreading and
causing negative sentiments, timely, proper, and effective risk communication is needed to help
ease peoples’ anxiety or negative attitudes regarding the COVID-19 pandemic especially
through social media.

Mental health is another issue that requires further investigations. Our chosen studies did not
address mental health issues based on age, as youth and adults could struggle with different
problems and varied interventions. Public health measures such as social distancing
implemented in the COVID-19 pandemic exacerbated risk factors and adverse health
behaviours at the individual and population levels. Studies have shown that social media data
were useful to detect mental health issues at the population level. Due to the early outbreak of
COVID-19 and the prevalence of social media (i.e. Weibo, WeChat) in China, two studies that
shed light on issues of mental health were both among the Chinese population. 78-79 Similar
deteriorating trend of mental health could happen in other regions. At the time of this writing,
British Columbia, Canada recorded the highest number of overdose deaths in May 2020.101

In terms of the surveillance of the COVID-19 pandemic, six chosen studies demonstrated
methods to detect or predict the number of COVID-19 cases using social media data. Unlike
other infectious diseases such as influenza and malaria, COVID-19 has not had real-time
monitoring surveillance developed with social media data according to our review. It is
possible that the pandemic has evolved so rapidly that finding COVID-19 vaccinations or
therapies has been prioritized over others. Besides, lacking accurate and reliable data sources
may discourage the development of the COVID-19 real-time surveillance. Moreover, it
remains unknown whether COVID-19 is just a one-time black swan event or would become
something like the seasonal flu. If the latter, then it may be meaningful and useful to establish
a real-time model to monitor the disease using social media data.

Government responses distributed via social media have increasingly played critical in
combating infodemics and promoting accurate and reliable information for the public.
However, little has been studied about how efficient and effective these official reactions lead
to belief or behavioural changes in real life. It also remained unknown that, compared with
infodemics, government posts would reach greater numbers of social media users or have
greater impacts on them.

YouTube has served as one of major platforms to spread information concerning the disease
control of COVID-19. Nonetheless, our chosen studies revealed undesirable quality for these
videos because they contained few recommended preventions from the governments or public
health organizations. It is a worrisome observation if accurate and reliable videos and other
types of information are not created and disseminated in a timely manner. Therefore, videos,
especially from public health authorities, should include accurate and reliable medical and
scientific information and use relevant hashtags to reach more audience, generate more views
and increase responses. Moreover, our selected studies were limited to YouTube videos only.
Additionally, a significant proportion of the studies were done using Sina Weibo, which,
although used by a large number of people, is exclusive of China, which may lead to an over-
representation of a single country in the scoping review.

In summary, although our scoping review has limitations embedded from the chosen studies,
it recognized six themes that have been studied so far, and future research directions are also
identified. Our adopted framework can serve as a fundamental and flexible guideline when
conducting studies related to social media and epidemiology.

Conclusions

Our review identified various topics, themes and methodological approaches in studies on
social media and COVID-19. Among the six identified themes, public attitudes comprised the
majority of the articles. Among the selected studies, Twitter was the leading social media
platform followed by Weibo. A minority of the studies included machine learning methods,
whereas a vast majority employed traditional statistical methods. Unlike influenza, we were
not able to find studies documenting real-time surveillance developed with social media data
on COVID-19. Our review also identified studies on COVID-19 related infodemics, mental
health and prediction. For COVID-19, accurate and reliable information through social media
platforms can play a crucial role in tackling infodemics, misinformation, and rumours.
Additionally, real time surveillance from social media about COVID-19 can be an important
tool in the armamentarium of interventions by public health agencies and organizations.

Contributors

HC and ZB conceived of the review. SFT conducted the screening, searched for articles,
performed data analysis, and wrote the first draft of the manuscript. TT refined the search
strategy and searched for articles. YY refined the search strategy and created the tables. LL
assisted in screening. ZB and HC supervised the review process and prepared the final draft for
submission. All authors contributed to the interpretation of results, manuscript preparation, and
revisions. All authors read and approved the final manuscript.

Declaration of interests

The authors declared no conflicts of interest

Acknowledgements

There was no funding for this review.

References
1) Medicine JHUa. Coronavirus Resource Center: World Map. © 2020 Johns Hopkins
University. 2020; Available at: https://coronavirus.jhu.edu/map.html.
2) Pérez-Escoda A, Jiménez-Narros C, Perlado-Lamo-de-Espinosa M, Pedrero-Esteban
LM. Social Networks' Engagement During the COVID-19 Pandemic in Spain: Health
Media vs. Healthcare Professionals. Int J Environ Res Public Health. 2020 Jul
21;17(14):5261.
3) Jordan SE, Hovet SE, Fung IC, Liang H, Fu K, Tse ZTH. Using Twitter for public
health surveillance from monitoring and prediction to public response. Data
2019;4(1):6.
4) Shah Z, Surian D, Dyda A, Coiera E, Mandl KD, Dunn AG. Automatically appraising
the credibility of vaccine-related web pages shared on social media: a Twitter
surveillance study. Journal of medical Internet research 2019;21(11):e14007.
5) Sinnenberg L, Buttenheim AM, Padrez K, Mancheno C, Ungar L, Merchant RM.
Twitter as a tool for health research: a systematic review. Am J Public Health
2017;107(1):e1-e8.
6) Steffens MS, Dunn AG, Wiley KE, Leask J. How organisations promoting
vaccination respond to misinformation on social media: a qualitative investigation.
BMC Public Health 2019;19(1):1-12.
7) Abd-Alrazaq A, Alhuwail D, Househ M, Hamdi M, Shah Z. Top concerns of tweeters
during the COVID-19 pandemic: infoveillance study. Journal of medical Internet
research 2020;22(4):e19016.
8) Freberg K, Palenchar MJ, Veil SR. Managing and sharing H1N1 crisis information
using social media bookmarking services. Public relations review 2013;39(3):178-
184.
9) Giustini D, Ali SM, Fraser M, Kamel Boulos MN. Effective uses of social media in
public health and medicine: a systematic review of systematic reviews. Online J
Public Health Inform. 2018 Sep 21;10(2):e215.
10) Al-Dmour H, Masa'deh R, Salman A, Abuhashesh M, Al-Dmour R. Influence of
Social Media Platforms on Public Health Protection Against the COVID-19 Pandemic
 via the Mediating Effects of Public Health Awareness and Behavioral Changes:
Integrated Model. J Med Internet Res. 2020 Aug 19;22(8):e19996.
11) Bridgman, A.; Merkley, E.; Loewen, PJ; Owen, T; Ruths, D.; Teichmann, L.; Zhilin,
O. (2020). The causes and consequences of COVID-19 misperceptions: understanding
the role of news and social media. The Harvard Kennedy School (HKS)
Misinformation Review
12) Tang L, Bie B, Park S, Zhi D. Social media and outbreaks of emerging infectious
diseases: A systematic review of literature. Am J Infect Control 2018;46(9):962-972.
13) Arksey H, O'Malley L. Scoping studies: towards a methodological framework.
International journal of social research methodology 2005;8(1):19-32.
14) Levac D, Colquhoun H, O'Brien KK. Scoping studies: advancing the methodology.
Implementation science 2010;5(1):69
15) Ose, S. O. (2016). Using Excel and Word to structure qualitative data. Journal of
Applied Social Science, 10(2), 147-162.
16) Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative
research in psychology, 3(2), 77-101.
17) Schillinger D, Chittamuru D, Ramírez AS. From “infodemics” to health promotion: A
novel framework for the role of social media in public health. American journal of
public health. 2020 Sep;110(9):1393-6.
18) World Health Organization. Novel Coronavirus(2019-nCoV) Situation Report - 13.
2020. Available at https://www.who.int/docs/default-source/coronaviruse/situation-
reports/20200202-sitrep-13-ncov-v3.pdf
19) Rovetta A, Bhagavathula AS. Global Infodemiology of COVID-19: Analysis of
Google Web Searches and Instagram Hashtags. J Med Internet Res
2020;22(8):e20673. doi:10.2196/20673.
20) Gallotti R, Valle F, Castaldo N, Sacco P, De Domenico M. Assessing the risks of
'infodemics' in response to COVID-19 epidemics. Nat Hum Behav (2020).
https://doi.org/10.1038/s41562-020-00994-6.
21) Pulido CM, Villarejo-Carballido B, Redondo-Sama G, Gómez A. COVID-19
infodemic: More retweets for science-based information on coronavirus than for false
information. International Sociology 2020:0268580920914755.
22) Kouzy R, Abi Jaoude J, Kraitem A, El Alam MB, Karam B, Adib E, et al.
Coronavirus Goes Viral: Quantifying the COVID-19 Misinformation Epidemic on
Twitter. Cureus 2020 March 13;12(3):e7255.
23) Moscadelli A, Albora G, Biamonte MA, Giorgetti D, Innocenzio M, Paoli S, Lorini C,
Bonanni P, Bonaccorsi B. Fake News and Covid-19 in Italy: Results of a Quantitative
Observational Study. Int. J. Environ. Res. Public Health 2020, 17(16), 5850;
https://doi.org/10.3390/ijerph17165850
24) Galhardi CP, Freire NP, Maria Minayo CS, Fagundes, MCM. Fact or Fake? An
analysis of disinformation regarding the Covid-19 pandemic in Brazil. Ciênc. saúde
coletiva [Internet]. 2020 Oct [cited 2020 Nov 18] ; 25( Suppl 2 ): 4201-4210.
Available from: http://www.scielo.br/scielo.php?script=sci_arttext&pid=S1413-
81232020006804201&lng=en. Epub Sep 30, 2020. https://doi.org/10.1590/1413-
812320202510.2.28922020.
25) Ahmed W, López Seguí F, Vidal-Alaball J, Katz MS. COVID-19 and the "Film Your
Hospital" Conspiracy Theory: Social Network Analysis of Twitter Data. J Med
Internet Res 2020;22(10):e22374. doi: 10.2196/22374.
26) Ahmed W, Vidal-Alaball J, Downing J, Seguí FL. COVID-19 and the 5G conspiracy
theory: social network analysis of Twitter data. Journal of Medical Internet Research.
2020;22(5):e19458.
27) Bruns, A, Harrington, S, Hurcombe, E. ‘Corona? 5G? or both?’: the dynamics of
COVID-19/5G conspiracy theories on Facebook. Media International Australia,
177(1), pp. 12–29. doi: 10.1177/1329878X20946113.
28) Islam MS, Sarkar T, Khan SH, Mostofa Kamal AH, Hasan SMM, Kabir A, Yeasmin
D, Islam MA, Amin Chowdhury KI, Anwar KS, Chughtai AA, Seale H. COVID-19-
Related Infodemic and Its Impact on Public Health: A Global Social Media Analysis.
The American Society of Tropical Medicine and Hygiene, 2020, 103(4): 1621 – 1629.
doi: https://doi.org/10.4269/ajtmh.20-0812.
29) Uyheng J, Carley KM. Bots and online hate during the COVID-19 pandemic: case
studies in the United States and the Philippines. J Comput Soc Sc (2020). doi:
https://doi.org/10.1007/s42001-020-00087-4
30) Brennen, J.S., Simon, F.M., Nielsen, R.K. Beyond (Mis)Representation: Visuals in
COVID-19 Misinformation. The International Journal of Press/politics. 2020 Oct.
31) Kwon, J, Grady, C, Feliciano, JT, Fodeh, SJ. Defining facets of social distancing
during the COVID-19 pandemic: Twitter analysis. Journal of Biomedical Informatics,
2020, 111. doi: https://doi.org/10.1016/j.jbi.2020.103601
32) Lai D, Wang D, Calvano J, Raja AS, He S. Addressing immediate public coronavirus
(COVID-19) concerns through social media: Utilizing Reddit's AMA as a framework
for Public Engagement with Science. PLOS ONE 15(10): e0240326.
https://doi.org/10.1371/journal.pone.0240326
33) Li Y, Twersky S, Ignace K, Zhao M, Purandare R, Bennett-Jones B, Weaver SR.
Constructing and Communicating COVID-19 Stigma on Twitter: A Content Analysis
of Tweets during the Early Stage of the COVID-19 Outbreak. Int. J. Environ. Res.
Public Health 2020, 17, 6847.
34) Medford RJ, Saleh SN, Sumarsono A, Perl TM, Lehmann CU. An "Infodemic":
Leveraging High-Volume Twitter Data to Understand Early Public Sentiment for the
Coronavirus Disease 2019 Outbreak. Open Forum Infectious Diseases, Volume 7,
Issue 7, July 2020, ofaa258, https://doi.org/10.1093/ofid/ofaa258
35) Al-Rawi A, Siddiqi M, Morgan R, Vandan N, Smith J, Wenham C. COVID-19 and
the gendered use of emojis on Twitter. J Med Internet Res 2020;22(11):e21646. doi:
10.2196/21646
36) Samuel, J, Mokhlesur Rahman GG, Ali N, Smaual Y, Pelaez A, Chong PHJ, Yakubov
M. Feeling Positive about Reopening? New Normal Scenarios from COVID-19 US
Reopen Sentiment Analytics. IEEE Access, vol. 8, pp. 142173-142190, 2020, doi:
10.1109/ACCESS.2020.3013933.
37) Odlum M, Cho H, Broadwell P, Davis N, Patrao M, Schauer D, Bales ME, Alcantara
C, Yoon S. Application of Topic Modeling to Tweets as the Foundation for Health
Disparity Research for COVID-19. Studies in Health Technology and Informatics,
2020;272: 24 – 27. doi: 10.3233/SHTI200484
38) Thelwall, M., Thelwall, S. COVID-19 tweeting in English: Gender differences.
Relaciones Públicas, 2020, 29(3). doi: https://doi.org/10.3145/epi.2020.may.01.
39) Xue J, Chen J, Chen C, Hu R, Zhu T. The Hidden Pandemic of Family Violence
During COVID-19: Unsupervised Learning of Tweets. J Med Internet Res
2020;22(11):e24361. doi: 10.2196/24361
40) Yigitcanlar T, Kankanamge N, Preston A, Gill PS, Rezayee M, Ostadnia M, Xia B,
Ioppolo G. How can social media analytics assist authorities in pandemic-related
policy decisions? Insights from Australian states and territories. Health Inf Sci Syst 8,
37 (2020). https://doi.org/10.1007/s13755-020-00121-9
41) Mohamad SM. Creative Production of 'COVID-19 Social Distancing' Narratives on
Social Media. Tijds. voor econ. en Soc. Geog., 111: 347-359. doi:
https://doi.org/10.1111/tesg.12430
42) Arpaci, I, Alshehabi, S, Al-Emran, M, Khasawneh, M, Mahariq, I, Abdeljawad, T,
Hassanien, AE. Analysis of twitter data using evolutionary clustering during the
COVID-19 pandemic. Computers, Materials & Continua, vol. 65, no.1, pp. 193–204,
2020.
43) Pastor CK. Sentiment Analysis of Filipinos and Effects of Extreme Community
Quarantine Due to Coronavirus (Covid-19) Pandemic. Available at SSRN 3574385
2020.
44) Boon-Itt S, Skunkan. Y. A text-mining analysis of public perceptions and topic
modeling during the COVID-19 pandemic using Twitter data.
JMIR Public Health Surveill 2020;6(4):e21978. doi: 10.2196/21978
45) Kurten S, Beullens K. #Coronavirus: Monitoring the Belgian Twitter Discourse on the
Severe Acute Respiratory Syndrome Coronavirus 2 Pandemic. Cyberpsychology,
Behavior, and Social Networking.ahead of print. doi;
http://doi.org/10.1089/cyber.2020.0341
46) De Santis, E, Martino, A, Rizzi, A. An Infoveillance System for Detecting and
Tracking Relevant Topics from Italian Tweets during the COVID-19 Event. IEEE
Access, vol. 8, pp. 132527-132538, 2020, doi: 10.1109/ACCESS.2020.3010033.
47) Essam BA, Abdo MS. How Do Arab Tweeters Perceive the COVID-19 Pandemic? J
Psycholinguist Res (2020). https://doi.org/10.1007/s10936-020-09715-6.
48) Lwin MO, Lu J, Sheldenkar A, Schulz PJ, Shin W, Gupta R, Yang Y. Global
Sentiments Surrounding the COVID-19 Pandemic on Twitter: Analysis of Twitter
Trends. JMIR Public Health Surveill 2020;6(2):e19447. doi: 10.2196/19447.
49) Chang A, Schulz PJ, Tu S, Liu MT. Blaming Devices in Online Communication of
the COVID-19 pandemic: Stigmatizing cues and negative sentiment gauged with
automated analytic techniques. Journal of Medical Internet Research.
02/10/2020:21504 (forthcoming/in press). doi: 10.2196/21504.
50) Das, S., Dutta, A. Characterizing public emotions and sentiments in COVID-19
environment: A case study of India. Journal of Human Behavior in the Social
Environment. doi: 10.1080/10911359.2020.1781015.
51) Chehal D, Gupta P, Gulati P. COVID-19 pandemic lockdown: An emotional health
perspective of Indians on Twitter. International Journal of Social Psychiatry. doi:
10.1177/0020764020940741.
52) Dheeraj, K. Analysing COVID-19 news impact on social media aggregation.
International Journal of Advanced Trends in Computer Science and Engineering.
2020, 9(3). doi: 10.30534/ijatcse/2020/56932020.
53) Park, HW, Park, S, Chong, M. Conversations and medical news frames on twitter:
Infodemiological study on COVID-19 in South Korea. J Med Internet Res
2020;22(5):e18897. doi: 10.2196/18897
54) Wicke P, Bolognesi MM. Framing COVID-19: How we conceptualize and discuss the
pandemic on Twitter. PLOS ONE 15(9): e0240010. doi:
https://doi.org/10.1371/journal.pone.0240010
55) Damiano, Amanda D, Catellier A, Jennifer R. A Content Analysis of Coronavirus
Tweets in the United States Just Prior to the Pandemic Declaration. Cyberpsychology,
Behavior, and Social Networking.ahead of print. doi:
http://doi.org/10.1089/cyber.2020.0425
56) Green, J, Edgerton. J, Naftel, D, Shoub, K, Cranmer, SJ. Elusive consensus:
Polarization in elite communication on the COVID-19 pandemic. Science Advances.
2020;6(28): eabc2717. doi: 10.1126/sciadv.abc2717.
57) Gozzi N, Tizzani M, Starnini M, Ciulla F, Paolotti D, Panisson A, Perra N. Collective
Response to Media Coverage of the COVID-19 Pandemic on Reddit and Wikipedia:
Mixed-Methods Analysis. J Med Internet Res 2020;22(10):e21597. doi:
10.2196/21597
58) Yu J, Lu Y, Muñoz-Justicia J. Analyzing Spanish News Frames on Twitter during
COVID-19-A Network Study of El País and El Mundo. Int. J. Environ. Res. Public
Health 2020, 17, 5414. doi: 10.3390/ijerph17155414.
59) Jelodar, H, Wang, Y, Orji, R, Huang, S. Deep Sentiment Classification and Topic
Discovery on Novel Coronavirus or COVID-19 Online Discussions: NLP Using
LSTM Recurrent Neural Network Approach. arXiv:2004.11695 [cs.IR]
60) Samuel J, Ali N, Mokhlesur Rahman GG, Esawi, E, Samuel, Y. COVID-19 public
sentiment insights and machine learning for tweets classification. Information 2020,
11, 314. doi: https://doi.org/10.3390/info11060314.
61) Chen Q, Min C, Zhang W, Wang G, Ma X, Evans R. Unpacking the black box: How
to promote citizen engagement through government social media during the COVID-
19 crisis. Comput Hum Behav 2020:106380.
62) Han X, Wang J, Zhang M, Wang X. Using Social Media to Mine and Analyze Public
Opinion Related to COVID-19 in China. International Journal of Environmental
Research and Public Health 2020;17(8):2788.
63) Zhao, Y, Cheng, S, Yu, X, Xu, H. Chinese public's attention to the COVID-19 epidemic
on social media: Observational descriptive study. J Med Internet Res
2020;22(5):e18825. doi: 10.2196/18825
64) Li J, Xu Q, Cuomo R, Purushothaman V, Mackey T. Data mining and content analysis
of the Chinese social media platform Weibo during the early COVID-19 outbreak:
retrospective observational infoveillance study. JMIR Public Health and Surveillance
2020;6(2):e18700.
65) Your Emotional Brain on Resentment, Part 2. Psych Central Professional. © 2020;
Available at: https://pro.psychcentral.com/your-emotional-brain-on-resentment-part-
2/.
66) Wang, T, Lu, K, Chow, KP, Zhu, Q. COVID-19 Sensing: Negative Sentiment Analysis
on Social Media in China via BERT Model. IEEE Access, vol. 8, pp. 138162-138169,
2020, doi: 10.1109/ACCESS.2020.3012595.
67) Xi, W., Xu, W., Zhang, X., & Ayalon, L. (2020). A thematic analysis of weibo topics
(Chinese twitter hashtags) regarding older adults during the COVID-19 outbreak. The
Journals of Gerontology: Series B.
68) Su, Y, Xue, J, Liu, X, Wu, P, Chen, J, Chen, C, Liu, T, Gong, W, Zhu, T. Examining
the Impact of COVID-19 Lockdown in Wuhan and Lombardy: A Psycholinguistic
Analysis on Weibo and Twitter. Int. J. Environ. Res. Public Health 2020, 17, 4552. doi:
https://doi.org/10.3390/ijerph17124552
69) Ma, R, Deng, Z, Wu, M. Effects of Health Information Dissemination on User Follows
and Likes during COVID-19 Outbreak in China: Data and Content Analysis. Int. J.
Environ. Res. Public Health 2020, 17, 5081. doi:
https://doi.org/10.3390/ijerph17145081
70) Xie T, Tan T, Li J. An Extensive Search Trends-Based Analysis of Public Attention on
Social Media in the Early Outbreak of COVID-19 in China. Risk Manag Healthc
Policy. 2020;13:1353-1364. doi: https://doi.org/10.2147/RMHP.S257473
71) Zhu B, Zheng X, Liu H, Li J, Wang P. Analysis of spatiotemporal characteristics of big
data on social media sentiment with COVID-19 epidemic topics. Chaos, Solitons &
Fractals. 2020;140. doi: https://doi.org/10.1016/j.chaos.2020.110123.
72) Kim B. Effects of social grooming on incivility in COVID-19. Cyberpsychology,
Behavior, and Social Networking 2020.
73) Budhwani, H, Sun, R. Exploring U.S. Shifts in Anti-Asian Sentiment with the
Emergence of COVID-19. Int. J. Environ. Res. Public Health 2020, 17, 7032.
74) Darling-Hammond S, Michaels EK, Allen AM, Chae DH, Thomas MD, Nguyen TT,
Mujahid MM, Johnson RC. After "The China Virus" Went Viral: Racially Charged
Coronavirus Coverage and Trends in Bias Against Asian Americans. Health Education
& Behavior. 2020;47(6):870-879. doi:10.1177/1090198120957949.
75) Nguyen TT, Criss S, Dwivedi P, Huang D, Keralis J, Hsu E, Phan L, Nguyen LH, Yardi
I, Glymour MM, Allen AM, Chae DH, Gee GC, Nguyen, QC. Exploring U.S. Shifts
in Anti-Asian Sentiment with the Emergence of COVID-19. Int. J. Environ. Res. Public
Health 2020, 17(19), 7032. doi: https://doi.org/10.3390/ijerph17197032
76) Jimenez‐Sotomayor MR, Gomez‐Moreno C, Soto‐Perez‐de‐Celis E. Coronavirus,
Ageism, and Twitter: An Evaluation of Tweets about Older Adults and COVID‐19.
Journal of the American Geriatrics Society. 2020 Apr 27.
77) Barrett AE, Michael C, Padavic I. Calculated Ageism: Generational Sacrifice as a
Response to the COVID-19 Pandemic. The Journals of Gerontology: Series B,
gbaa132, doi: https://doi.org/10.1093/geronb/gbaa132.
78) Gao J, Zheng P, Jia Y, Chen H, Mao Y, Chen S, et al. Mental health problems and
social media exposure during COVID-19 outbreak. Plos one 2020;15(4):e0231924.
79) Li S, Wang Y, Xue J, Zhao N, Zhu T. The impact of COVID-19 epidemic declaration
on psychological consequences: a study on active Weibo users. International journal of
environmental research and public health 2020;17(6):2032.
80) de Vreese, CH, Neijens P. Measuring media exposure in a changing communications
environment. Journal of communication methods and measures. 2016;10(2-3):69-80.
81) Qin L, Sun Q, Wang Y, Wu K, Chen M, Shia B, et al. Prediction of number of cases of
2019 novel coronavirus (COVID-19) using social media search index. International
journal of environmental research and public health 2020;17(7):2365.
82) Li C, Chen LJ, Chen X, Zhang M, Pang CP, Chen H. Retrospective analysis of the
possibility of predicting the COVID-19 outbreak from Internet searches and social
media data, China, 2020. Eurosurveillance 2020;25(10):2000199.
83) Peng Z, Wang R, Liu L, Wu H. Exploring urban spatial features of COVID-19
transmission in Wuhan based on social media data. ISPRS Int. J. Geo-Inf. 2020, 9, 402.
doi: https://doi.org/10.3390/ijgi9060402
84) Liu D, Wang Y, Wang J, Liu J, Yue Y, Liu W, Zhang F, Wang Z. Characteristics and
Outcomes of a Sample of Patients With COVID-19 Identified Through Social Media
in Wuhan, China: Observational Study. J Med Internet Res 2020;22(8):e20108. doi:
10.2196/20108.
85) O'Leary, DE, Storey, VC. A Google–Wikipedia–Twitter Model as a Leading Indicator
of the Numbers of Coronavirus Deaths. Intell Sys Acc Fin Mgmt. 2020; 27: 151– 158.
doi: https://doi.org/10.1002/isaf.1482
86) Zhu Y, Fu KW, Grépin KA, Liang H, Fung IC. Limited Early Warnings and Public
Attention to Coronavirus Disease 2019 in China, January–February, 2020: A
Longitudinal Cohort of Randomly Sampled Weibo Users. Disaster Medicine and Public
Health Preparedness. 2020 Apr 3:1-4.
87) Li L, Zhang Q, Wang X, Zhang J, Wang T, Gao T, et al. Characterizing the propagation
of situational information in social media during covid-19 epidemic: A case study on
weibo. IEEE Transactions on Computational Social Systems 2020;7(2):556-562.
88) Rufai SR, Bunce C. World leaders’ usage of Twitter in response to the COVID-19
pandemic: a content analysis. Journal of Public Health 2020.
89) Merkley E, Bridgman A, Loewen PJ, Owen T, Ruths D, Zhilin O. A Rare Moment of
Cross-Partisan Consensus: Elite and Public Response to the COVID-19 Pandemic in
Canada. Canadian Journal of Political Science/Revue canadienne de science politique
2020:1-8.
90) Netizen. Wikipedia. © 2020; Available at: https://en.wikipedia.org/wiki/Netizen.
91) Sutton J, Renshaw SL, Butts CT. COVID-19: Retransmission of official
communications in an emerging pandemic. PLOS ONE 15(9): e0238491. doi:
https://doi.org/10.1371/journal.pone.0238491.
92) Wang, Y, Hao, H, Platt, LS. Examining risk and crisis communications of government
agencies and stakeholders during early-stages of COVID-19 on Twitter. Computers in
Human Behavior. 114. doi: https://doi.org/10.1016/j.chb.2020.106568
93) Basch CH, Hillyer GC, Meleo-Erwin ZC, Jaime C, Mohlman J, Basch CE. Preventive
behaviors conveyed on YouTube to mitigate transmission of COVID-19: cross-
sectional study. JMIR public health and surveillance 2020;6(2):e18807.
94) Basch CE, Basch CH, Hillyer GC, Jaime C. The role of YouTube and the
entertainment industry in saving lives by educating and mobilizing the public to adopt
behaviors for community mitigation of COVID-19: successive sampling design study.
JMIR Public Health and Surveillance 2020;6(2):e19145.
95) Hakimi AA, Armstrong WB. Hand Sanitizer in a Pandemic: Wrong Formulations in
the Wrong Hands. The Journal of Emergency Medicine. 2020. Doi:
https://doi.org/10.1016/j.jemermed.2020.07.018.
96) Khatri P, Singh S, Belani NK, Leng YY, Lohan R, Wei LY, et al. YouTube as source
of information on COVID-19 outbreak: a cross sectional study of English and Mandarin
content. Travel medicine and infectious disease 2020:101636.
97) Hernández-García I, Giménez-Júlvez T. Characteristics of YouTube Videos in
Spanish on How to Prevent COVID-19. Int. J. Environ. Res. Public Health 2020,
17(13), 4671. doi: https://doi.org/10.3390/ijerph17134671.
98) Moon, H, Lee, GH. Evaluation of Korean-language COVID-19-related medical
information on YouTube: Cross-sectional infodemiology study. J Med Internet Res
2020;22(8):e20775. doi: 10.2196/20775
99) Ozdede M, Peker I. Analysis of Dentistry YouTube Videos Related To COVID-19.
Braz. Dent. J. [Internet]. 2020 Aug [cited 2020 Nov 18] ; 31( 4 ): 392-398.
Available from: http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0103-
64402020000400392&lng=en. Epub Sep 04, 2020. http://dx.doi.org/10.1590/0103-
6440202003767.
100) Yüce MÖ, Adalı E, Kanmaz B. An analysis of YouTube videos as educational
resources for dental practitioners to prevent the spread of COVID-19. Ir J Med Sci
(2020). doi: https://doi.org/10.1007/s11845-020-02312-5
101) Rhianna Schmunk. B.C. records highest number of fatal overdoses in a single
month, with 170 deaths. 2020; Available at: https://www.cbc.ca/news/canada/british-
columbia/overdose-deaths-bc-1.5607792.
Author(s) Yea Origin Aim/Them Social Study Population / Methods Key Findings
r e Media Sample Size
Liu D, Wang 2020 China Detection Weibo Weibo messages • Gathered data via • Older age, diffuse distribution, and
Y, Wang J, or between Jan 20 – Feb 15, Weibo, then hypoxemia are factors that can
Liu J, Yue Y, prediction 599 patients participated followed up with help clinicians identify patients
Liu W, Zhang of COVID- telephone with COVID-19 who have poor
F, Wang Z. 19 cases • Statistical analysis prognosis
taken with Fisher • Aggregated data from social media
exact test can also be comprehensive,
• Rates of death by immediate, and informative in
Kaplan-Meier disease prognosis.
method
• Used multivariate
Cox regression to
determine risk
factor for mortality
O'Leary, DE, 2020 USA Detection Google, Google Trend searches • Regression analysis • To model of the number of cases,
Storey, VC. or Wikipedia, for “coronavirus” and the current Wikipedia page views,
prediction Twitter “COVID‐19” started on 1‐week‐ahead tweets, and 2‐week
of COVID- 21 January 2020 to 5 ahead Google searches were
19 cases April 2020. Wikipedia employed.
page views for • To model of the number of deaths,
“coronavirus” and each variable was placed 1 week
“COVID‐19” started on further ahead.
12 January 2020 to 5
April 2020. The number
of Twitter original
tweets, starting on 27
January 2020 to 5 April
2020. The numbers of
COVID-19 cases and
deaths in the USA from
www.worldometers.info

Peng Z, Wang 2020 China Detection Weibo 1,200 records • Spatiotemporal • Elderlies are the group of high-risk
R, Liu L, Wu or distribution of and high-prevalence in the COVID-
H. prediction COVID-19 cases in 19 outbreak, account for > 50% of
 of COVID- the main urban area the total number of Weibo help
19 cases of Wuhan. seekers.
• Kernel Density • Early transmission of COVID-19 in
Analysis. Wuhan could be divide into three
• Ordinary Least phrases: Scattered infection,
Square Regression community spread, and full-scale
outbreak.
Cuilian Li, Li 2020 China Disease Google Keywords "coronavirus" • Lag correlation • Lag correlations showed a max
Jia Chen, prediction Trends, and "pneumonia" correlation at 8-12 days for lab-
Xueyu Chen, Baidu searched, and trend data confirmed cases & 6-8 days for
Mingzhi Index, Sina collected from Google suspected cases.
Zhang, Chi Weibo Trends, Baidu Index, &
Pui Pang, Index Sina Weibo Index from
Haoyu Chen Jan 2 to Feb 20, 2020
Lei Qin, 2020 China Disease Baidu SMSI for dry cough, • Subset selection • New suspected COVID-19 case
Qiang Sun, prediction search fever, chest distress, • Forward selection. numbers correlated significantly
Yidan Wang, index (BSI) coronavirus, and • Lasso regression. with the lagged series of SMSI.
Ke Fei Wu, pneumonia from Dec 31, • Ridge regression. • SMSI could detected 6 - 9 days
Mingchih 2019 to Feb 9, 2020. • Elastic net earlier than new suspected cases.
Chen, Ben New suspected cases of
Chang Shia, COVID-19 data from Jan
Szu Yuan Wu 20 to Feb 9, 2020.
Yuner Zhu, 2020 China Disease Sina Weibo 1,101 COVID-19 related • Descriptive • Attention to COVID-19 was
King-Wa Fu, prediction Weibo posts from Dec statistics: numbers, limited until China openly admitted
Karen A 31, 2019 to Feb 12, 2020 percentage human-to-human transmission on
Grépin, Hai • Time series Jan 20, 2020.
Liang, Isaac analysis • Attention quickly increased and
Chun-Hai remained high over time.
Fung
Charles E 2020 United Governme YouTube 100 most widely viewed • Descriptive • < 50% the videos in either sample
Basch, Corey States nt YouTube videos as of analysis: frequency, covered any of the prevention
H Basch, responses Jan 31 & Mar 20, 2020 percent, mean, behaviours recommended by the
Grace C with keywords standard deviation US CDC.
Hillyer, "Coronavirus" in
Christie Jaime English, with English
subtitles, or in Spanish
Eric Merkley, 2020 Canada Governme Twitter, 33,142 tweets from 292 • Linear regression • No MPs from any party
Aengus nt Google social media accounts of downplaying the pandemic.
Bridgman, responses Trends federal MPs from Jan 1 • No association between
Peter John to Mar 28, 2020. 87 Conservative Party vote share &
Loewen, Google search trends for Google search interest in the
Taylor Owen, the search term coronavirus.
Derek Ruths, "coronavirus" in the 1st
Oleg Zhilin half (1-14) and 2nd half
(15-31) of Mar,
2020. A survey of 2,499
Canadian citizens ≥ 18 yr
from Apr 2 to 6, 2020
Lifang Li, 2020 China Governme Sina Weibo 36,746 Weibo data from • Linear regression. • Classify theCOVID-19-related
Qingpeng nt Dec 30, 2019 to Feb 01, SVM. NB. NLP information into 7 types of
Zhang, Xiao responses 2020; a random sample situational information & their
Wang, Jun of 3,000 Weibo as predictors.
Zhang, Tao training dataset
Wang, Tian-
Lu Gao, Wei
Duan, Kelvin
Kam-fai Tsoi,
Fei-Yue Wang
Priyanka 2020 Governme Youtube 150 videos on Feb 1, 2 • Descriptive • Mean DISCERN score for
Khatri, nt for “2019 novel analysis: percent, reliability was 3.12/5 for English
Shweta R responses coronavirus” (50), mean and 3.25/5 for Mandarin videos.
Singh, Neeta “Wuhan virus” in • DISCERN score • Mean cumulative MICI score of
Kesu Belani, English (50), and • MICI score useful videos was 6.71/25 for
Yin Leng Mandaring (50) English & 6.28/25 for Mandarin.
Yeong, Rahul
Lohan, Yee
Wei Lim,
Winnie ZY
Teo
Sohaib R 2020 United Governme Twitter 203 viral tweets from G7 • Qualitative design - • 82.8% were “informative.”
Rufai, Catey States nt world leaders from Nov content analysis • 9.4% were “morale-boosting.”
Bunce responses 17, 2019 to Mar 17, 2020 • 6.9% were “political.”
with keywords "COVID-
 19" or "coronavirus" and
a minimum of 500
"likes."
Sutton J, 2020 USA Governme Twitter 690 accounts • Chi-square • Systematic changes in message
Renshaw SL, nt representing public analyses strategies over time and identify
Butts CT. responses health, emergency • Negative binomial key features that affect message
management and elected regression passing, both positively and
officials and 149,335 modeling negatively.
tweets • The results have the potential to aid
in message design strategies as the
pandemic continues, or in similar
future events.
Wang, Y, 2020 USA Governme Twitter 13,598 COVID-19 • Temporal analysis • 16 categories of message types
Hao, H, Platt, nt related tweets from • Networking were manually annotated.
LS. responses January 1, 2020 to April analysis • Inconsistencies and incongruencies
27, 2020 were identified in four critical
topics—wearing masks,
assessment of risks, stay at home
order, and disinfectant and
sanitizer.
• Network analysis revealed
increased communication
coordination over time.
Ahmed W, 2020 UK Infodemics Twitter 22,785 tweets and 11,333 • Social network • The most important drivers of the
López Seguí Twitter users with analysis conspiracy theory are ordinary
F, Vidal- #FilmYourHospital • User analysis citizens.
Alaball J, hashtag April 13-20, • YouTube was the information
Katz MS. 2020 source most linked to by users.
The most retweeted post belonged
to a verified Twitter user.
Brennen, J.S., 2020 UK Infodemics Digital 96 samples of visuals • Qualitative coding • Organized all findings into six
Simon, F.M., visual from January 2020 trends – authoritative agency,
Nielsen, R.K. media through March 2020 virulence, medical efficacy,
intolerance, prophecy, satire
• A small number of manipulated
visuals, all were produced using
simple tools; there were no
 examples of “deepfakes” or other
artificial intelligence-based
techniques.
Bruns, A, 2020 Australia Infodemics Facebook 89,664 distinct Facebook • Time series. • Significantly greater volume of 5G
Harrington, S, posts from 1 January to Network analysis rumour posts on Facebook after
Hurcombe, E. 12 April 2020 March 19, 2020.
• Network analysis revealed that
coalition of various groups were
brought together by conspiracy
theories about COVID-19 and 5G
technology.
Cristina M 2020 Spain Infodemics Twitter 942 valid tweets between • Communicative • Misinformation tweeted more but
Pulido, Feb 6 & 7, 2020 content analysis retweeted less than science-based
Beatriz evidence tweets.
Villarejo- • Science-based evidence tweets had
Carballido, more engagement than
Gisela misinformation.
Redondo-
Sama, Aitor
Gómez
Galhardi CP, 2020 Brazil Infodemics WhatsApp, Fake news collected • Quantitative • WhatsApp is the primary channel
Freire NP, Instagram, from March 17 to April content analysis for sharing fake news, followed by
Maria Minayo Facebook 10, 2020 based on the Eu Instagram and Facebook.
CS, Fagundes, Fiscalizo app’s data
MCM.
Gallotti R, 2020 Italy Infodemics Twitter > 100 million Tweets • Developed an • Before the rise of COVID-19
Valle F, Infodemic Risk cases, entire countries expose
Castaldo N, Index measurable waves of potentially
Sacco P, De unreliable information, posing a
Domenico M. serious threat to public health.
Islam MS, 2020 Bangladesh Infodemics Online 2,311 reports related to • Descriptive • Misinformation fueled by rumors,
Sarkar T, platforms, COVID-19 related analysis stigma, and conspiracy theories
Khan SH, websites, infodemic reports • Spatial distribution can have potentially severe
Mostofa Facebook, between Dec 31, 2019 – analysis implications on public health if
Kamal AH, Twitter, April 5, 2020 prioritized over scientific
Hasan SMM, online guidelines.
Kabir A, papers
Yeasmin D, • Governments and other agencies
Islam MA, must understand the patterns of
Amin COVID-19–related rumors,
Chowdhury stigma, and conspiracy theories
KI, Anwar circulating the globe so that they
KS, Chughtai can develop appropriate risk
AA, Seale H. communication messages.
Moscadelli A, 2020 Italy Infodemics “fake 2102 articles between • Social media trend • Links containing “fake news” were
Albora G, news” and December 31, 2019 and analysis using the shared 2,352,585 times, accounting
Biamonte correspondi April 30, 2020 BuzzSumo for 23.1% of all reviewed articles.
MA, Giorgetti ng verified Application
D, Innocenzio news
M, Paoli S, circulated
Lorini C, in Italy
Bonanni P,
Bonaccorsi B.
Ramez 2020 Lebanon Infodemics Twitter 673 English tweets • Descriptive • 153 tweets (24.8%) had
Kouzy, Joseph collected on Feb 27, statistics misinformation.
Abi Jaoude, 2020 • Bar chart • 107 (17.4%) had unverifiable
Afif Kraitem, • Chi-square statistic information.
Molly B El to calculate p-value • Misinformation rate higher in
Alam, Basil (2-sided; 0.05) for informal individual/group accounts
Karam, Elio the association (33.8%)
Adib, Jabra between
Zarka, Cindy account/tweet
Traboulsi, characteristics and
Elie W Akl, the presence of
Khalil misinformation or
Baddour unverifiable
information about
COVID-19
Rovetta A, 2020 Italy Infodemics Google, 2 million Google Trends • Classification of • Globally, there is a growing
Bhagavathula Instagram queries and Instagram infodemic interest in COVID-19, and
AS. hashtags February 20, monikers numerous infodemic monikers
2020, to May 6, 2020 • Computed the continue to circulate on the
average peak internet.
 volume with a 95%
CI
Wasim 2020 United Infodemics Twitter 10,140 tweets collected • Descriptive • 34.8% linked 5G & COVID-19.
Ahmed, Josep Kingdom from 19:44 UTC on statistics: numbers, • 32.2% denounced the conspiracy
Vidal-Alaball, Friday, Mar 27, 2020 to percent theory.
Joseph 10:38 UTC on Saturday, • Social network
Downing, Apr 04, 2020 analysis
Francesc • Content analysis
López Seguí
Uyheng J, 2020 USA, Infodemics Twitter 12 million tweets from • Hate speech scored • Analysis reveals idiosyncratic
Carley KM. Philippines 1.6 million users from assigned to each relationships between bots and
the US and 15million tweet using ML hate speech across datasets,
tweets from 1 million algorithm highlighting different network
users from the • Bot scores were dynamics of racially charged
Philippines between Mar assigned to each toxicity in the US and political
5 and Mar 19, 2020 user via BotHunter conflicts in the Philippines.
algorithm • Bot activity is linked to higher hate
• Social media in both countries, especially in
analysis via ORA communities which are denser and
software more isolated from others.
• Network analysis
via centrality
analysis
• Cluster analysis via
Leiden algorithm
Junling Gao, 2020 China Mental Sina Weibo Online survey on • Multivariable • Frequently SME positively
Pinpin Zheng, Health Wenjuanxing platform logistic regression associated with high odds of
Yingnan Jia, from Jan 31 to Feb 2, anxiety (OR = 1.72, 95% CI: 1.31 -
Hao Chen, 2020; with 4,872 2.26) & CDA (OR = 1.91, 95%
Yimeng Mao, Chinese citizens ≥ 18 yr CI: 1.52 - 2.41)
Suhong Chen, from 31 provinces and
Yi Wang, Hua autonomous regions in
Fu, Junming China
Dai
Sijia Li, Yilin 2020 China Mental Sina Weibo Weibo posts from 17,865 • Sentiment analysis • Negative emotions & sensitivity to
Wang, Jia Health active Weibo users • Paired sample t-test social risks increased.
Xue, Nan
Zhao, between 13 Jan to 26 • Scores of positive emotions & life
Tingshao Zhu Jan, 2020 satisfactions decreased.
Hakimi AA, 2020 USA Prevention YouTube The first 100 videos on • Codified video • Most of the videos failed to
Armstrong Education YouTube with the most content describe labeling storage
WB. in Videos views using the search • Assessed using containers, 69% of videos
term “DIY hand Cohen kapa encouraged the use of oils or
sanitizer.” • Descriptive perfumes to enhance hand sanitizer
statistics calculated scent, and 2% of videos promoted
• Assessed by Chi- the use of coloring agents to be
square test with 2- more attractive for use among
sided p-value <0.5 children specifically.
• Significantly increased average
number of daily calls to poison
control centers regarding unsafe
pediatric hand sanitizer exposure
since the first confirmed COVID-
19 patient in the US.

• Significantly increased average

number of daily calls in March
2020 compared with the previous 2
years.
Hernández- 2020 Spain Prevention YouTube 129 videos in Spanish • Univariate analysis • Information from YouTube in
García I, Education with the terms • Multiple logistic Spanish on basic measures to
Giménez- in Videos “prevencion regression model prevent COVID-19 is usually not
Júlvez T. coronavirus” and very complete and differs
“prevencion COVID19” according to the type of
authorship.
Moon, H, Lee, 2020 South Prevention YouTube 105 the most viewed • Modified • 37.14% (39/105) contained
GH. Korea Education YouTube videos from DISCERN index misleading information.
in Videos January 1, 2020, to April • JAMAS benchmark • Independent user–generated videos
30, 2020 criteria. showed the highest proportion of
• GQS. misleading information at 68.09%
• TCCI. (32/47).
• MICI • Misleading videos had more likes,
fewer comments, and longer
running times than useful videos.
 • “Transmission and precautionary
measures” were the most
commonly covered content.
Ozdede M, 2020 Turkey Prevention YouTube The top 116 English- • Precision indices • High number of views on dentistry
Peker I. Education language videos with at and total video YouTube videos related to
in Videos least 300 views information and COVID-19.
quality index • The quality and usefulness of these
(VIQI) scores were videos are moderate.
calculated
Yüce MÖ, 2020 Turkey Prevention YouTube 55 English videos about • Modified • Only 2 videos (3.6%) were found
Adalı E, Education COVID-19 control DISCERN to be of good quality, while 24
Kanmaz B. in Videos procedures for dental instrument videos (43.6%) were of poor
practice on 31 March • Descriptive quality.
2020 between 9 AM and statistics
6 PM
Al-Rawi A, 2020 Canada Public Twitter Over 50 million tweets • Mixed method - • Identified 5 major themes in the
Siddiqi M, Attitudes referencing the hashtags Analyzed emoji use analysis: morbidity fears, health
Morgan R, #Covid-19 and #Covid19 by each gender concerns, employment and
Vandan N, for more than 2 months category financial issues, praise for frontline
Smith J, in early 2020 • The top 600 emojis workers, and unique gendered
Wenham C. were manually emoji use.
classified based on • Most emojis are extremely positive
their sentiment by gender, but discussions about
gender and minorities (female to a
lesser extent) are more negative
than male.
• When discussing certain topics, in
addition to discussions between
men and women, there are many
differences
• Use several unique gender emojis
to express specific issues
Alaa Abd- 2020 Qatar Public Twitter 2.8 English million • Word frequencies • Identified 12 topics and grouped
Alrazaq, Dari Attitudes tweets (167,073 unique of single into 4 themes.
Alhuwail, tweets from 160,829 (unigrams) and • Mean sentiment positive for 10
Mowafa unique users) from Feb. double words topics & negative for 2 topics.
Househ, 2 to Mar 15, 2020 (bigrams)
Mounir • Sentiment analysis
Hamdi, Zubair • Average number of
Shah retweets, likes, and
followers for each
topic
• Interaction rate per
topic. LDA for
topic modelling
Arpaci, I, 2020 Turkey Public Twitter 43 million tweets • Evolutionary • Unigram terms appear more
Alshehabi, S, Attitudes between March 22 and clustering analysis frequently than bigram and trigram
Al-Emran, M, March 30, 2020 terms.
Khasawneh, • During the epidemic, a large
M, Mahariq, I, number of tweets about COVID-19
Abdeljawad, were distributed and attracted
T, Hassanien, widespread public attention.
AE. • High-frequency words e.g.
“death”, “test”, “spread”, and
“lockdown” indicate that people
are afraid of being infected, and
the affected are fear of death
• People agree to stay at home due
to fear of spread and have called
for social distancing since they
become aware of COVID-19.
Barrett AE, 2020 USA Public Twitter 188 tweets about • Thematic analysis • 90% of all tweets opposed
Michael C, Attitudes Governor Dan Patrick’s calculated ageism, while only 5%
Padavic I. Mar 23 statement on supported it and 5% conveyed no
generational self- position.
sacrifice. • Opposition centered on moral
critiques, political-economic
critiques, assertions of older
adults’ worth, and public health
arguments. Support centered on
individual responsibility and
patriotism.
Boon-Itt S, 2020 Thailand Public Twitter 107,990 English tweets • Sentiment analysis • Sentiment analysis showed that a
Skunkan. Y. Attitudes related to COVID-19 • Topic modeling predominantly negative feeling
between December 13, using LDA toward COVID-19 pandemic.
2019 and March 9, 2020 • Topic modeling revealed 3 themes
relating to COVID-19 and the
outbreak: the COVID-19 pandemic
emergency, how to control
COVID-19, and reports on
COVID-19.
Budhwani, H., 2020 USA Public Twitter 16,535 “Chinese virus” • Descriptive • Nearly 10-fold increase at the
Sun, R. Attitudes or “China virus” tweets analysis national level.
between March 9 and • Spatial analysis • All 50 states witnessed an increase
March 15, 2020, 177,327 in the number of tweets
tweets between March 19 exclusively mentioning “Chinese
and March 25, 2020. virus” or “China virus” instead of
coronavirus disease (COVID-19)
or coronavirus.
• On average, 0.38 tweets
referencing “Chinese virus” or
“China virus” were posted per
10,000 people at the state level in
the preperiod, and 4.08 of these
stigmatizing tweets were posted in
the postperiod, also indicating a
10-fold increase.
Bumsoo Kim 2020 South Public Twitter 27,849 individual tweets • Binary logistic • Social network size was a negative
Korea Attitudes about COVID-19 regression predictor of incivility.
between Feb 10 and 14, • Semantic network
2020 analysis
Chang A, 2020 Taiwan Public 10 news 1.07 million Chinese • Deductive analysis • Online news served as a hotbed for
Schulz PJ, Tu Attitudes websites, texts from December 30, negativity and for driving
S, Liu MT. 11 2019, to March 31, 2020 emotional social posts.
discussion • Stigmatizing language linked to
forums, 1 the COVID-19 pandemic shows a
social lack of civic responsibility that
network, 2 encourages bias, hostility, and
principal discrimination.
 media
sharing
networks
Chehal D, 2020 India Public Twitter 29,554 tweets of • Sentiment analysis • A positive approach in lockdown
Gupta P, Attitudes lockdown 2.0; 47,672 using NRC 2.0 whereas a negative approach in
Gulati P. tweets of lockdown 3.0 Emotion Lexicon lockdown 3.0.
Cherish Kay 2020 Philippines Public Twitter Tweets were collected on • NLP for sentiment • Negative sentiments increased
Pastor Attitudes 3 Tuesdays in Mar 2020 analysis over time.
since lockdown in
Philippines
Damiano, 2020 USA Public Twitter 600 English tweets from • Frequencies • Neutral sentiment
Amanda D, Attitudes the USA were selected: • Chi-square • Tweets about COVID-19 risks and
Catellier A, 300 from February and statistics emotional outrage accounted <
Jennifer R 300 from March 2020. 50%.
• Few tweets related to blame.
Darling- 2020 USA Public Twitter 339,063 tweets from • Local polynomial • Implicit Americanness Bias
Hammond S, Attitudes non-Asian respondents regression steadily decreased from 2007 to
Michaels EK, of the Project Implicit • Interrupted time- 2020.
Allen AM, “Asian IAT” from 2007– series analyses • When media entities began using
Chae DH, 2020 and were broken stigmatizing terms like “Chinese
Thomas MD, into 2 datasets. The first virus” starting from March 8,
Nguyen TT, data set from January 1, 2020, Implicit Americanness Bias
Mujahid MM, 2007, to February 10, began to increase.
Johnson RC. 2020. The second data • Such bias was more pronounced
set begins following the among conservative individuals.
February 11 to March 31,
2020.
Das, S., Dutta, 2020 India Public Twitter 410,643 tweets with • NRC lexicon for • For the broad corpus-level
A. Attitudes #IndiaLockdown and corpus-level analysis, the context of
#IndiafightsCorona from emotion mining positiveness is significantly higher
Mar 22-April 21, 2020. • R’s sentiment to than negative sentiments.
create additional • However, the positive sentiment
sentiment scores. trends are not so different from the
LDA for topic negative sentiment trends when the
models analysis was performed at
individual tweet level.
 • nltk to develop • The results show that the
sentiment-based discussion of “COVID-19 in
topic models India” on Twitter contains slightly
more positive sentiments than
negative sentiments.
De Santis, E, 2020 Italy Public Twitter 1,044,645 tweets • A general purpose • Energy Evolution Through Time
Martino, A, Attitudes methodological was monitored
Rizzi, A. framework, • Daily hot topics were identified.
grounded on a
biological
metaphor and on a
chain of NLP and
graph analysis
techniques
Dheeraj, K. 2020 India Public Reddit 868 related COVID-19 • Fetching the • On reddit related to COVID-19
Attitudes postings articles: PRAW articles, 50% neutral, 22%
wrapper. Data positive, and 28% negative
preprocessing:
Reddit API and
NLTK library
Essam BA, 2020 Egypt Public Twitter 1,920,593 tweets with • Thematic analysis • The dominant themes that were
Abdo MS. Attitudes Corona, Coronavirus, closely related to coronavirus
COVID19 keywords tweets included the outbreak of the
from Feb 1 – April 30, pandemic, metaphysics responses,
2020 signs and symptoms in confirmed
cases, and conspiracies.
• The psycholinguistic analysis also
showed that tweeters maintained
high levels of affective talk, which
was loaded with negative emotions
and sadness.
• LIWC’s psychological categories
of religion and health dominated
the Arabic tweets discussing the
pandemic situation."
Fu Lian Yin, 2020 China Public Sina Weibo Weibo posts from Dec • M-SDL model • Model reproduction ratio declined
Jia Hui Lv, Attitudes 31, 2019 to Feb 7, 2020 from 1.78
Xiao Jian to 0.97, showing the peak of public
Zhang, Xin opinion passed but it’d continue.
Yu Xia, Jian
Hong Wu
Gozzi N, 2020 Italy, UK, Public News, 227,768 web-based news • Linear regression. • Collective attention was mainly
Tizzani M, USA, & Attitudes YouTube, articles from February 7 Topic modeling driven by media coverage rather
Starnini M, Canada Reddit, to May 15, 2020. 13,448 using LDA than epidemic progression, rapidly
Ciulla F, Wikipedia YouTube videos from became saturated, and decreased
Paolotti D, February 7 to May 15, despite media coverage and
Panisson A, 2020. 107,898 English COVID-19 incidence remaining
Perra N. user posts and 3,829,309 high.
comments on Reddit • Reddit users were generally more
from February 15 to May interested in health, data regarding
15, 2020. 278,456,892 the new disease, and interventions
views of COVID-19– needed to halt the spreading with
related Wikipedia pages respect to media exposure.
from February 7 to May
15, 2020
Green, J, 2020 USA Public Twitter 19,803 tweets from • RF • Democrats discussed the crisis
Edgerton. J, Attitudes Democrats and 11,084 more frequently–emphasizing
Naftel, D, tweets from Republicans threats to public health and
Shoub, K, between 17 January to 31 American workers–while
Cranmer, SJ. March 2020. Republicans placed greater
emphasis on China and businesses.
Jelodar, H, 2020 China Public Reddit 563,079 COVID-19– • Topic modeling • The results showed a novel
Wang, Y, Attitudes related English using LDA & application for NLP based on an
Orji, R, comments from Reddit PLSA LSTM model to detect meaningful
Huang, S. between January 20, • Sentiment latent-topics and sentiment-
2020 and March 19, classification using comment-classification on
2020. RNN COVID-19–related issues on
social media.
Jiawei Li, 2020 China Public Sina Weibo 115,299 Weibo posts • Linear regression • Positive correlation between the
Qing Xu, attitudes from Dec 23, 2019 to Jan model number of Weibo posts & number
Raphael 30, 2020; 11,893 of them • Qualitative content of reported cases, with 10 more
Cuomo, Vidya collected from Dec 31, analysis COVID-19 cases per 40 posts
Purushothama 2019 to Jan 20, 2020 for • Posts grouped into 4 themes.
qualitative analysis.
n, Tim Total daily cases of
Mackey COVID-19 in Wuhan
were obtained from the
Chinese National Health
Commission
Jimenez- 2020 Mexico Public Twitter A random sample of 351 • Qualitative content • The most common types of tweets
Sotomayor, Attitudes out of 18,128 tweets classification were personal opinions (31.9%),
MR, Gomez- from March 12 and followed by informative tweets
Moreno, C, March21, 2020 were (29.6%), jokes/ridicule (14.3%),
Soto-Perez- analyzed. and personal accounts (13.4%).
de-Celis, E. • 72 tweets (21.9%) likely intended
to ridicule or offend someone and
21.1% had content implying that
the life of older adults was less
valuable or downplayed the
relevance of COVID-19.
Kurten S, 2020 Belgian Public Twitter 373,908 tweets and • Time series • Significant COVID-19 events
Beullens K. Attitudes retweets from February • Network bigrams immediately increased the number
25, 2020 to March 30, • Emotion lexicon of tweets.
2020. • LDA • Most topics focused on the need of
EU collaboration to tackle the
pandemic.
Kwon, J, 2020 USA Public Twitter 259,529 unique tweets • Trending analysis • Early facets of social distancing
Grady, C, Attitudes containing “coronavirus” • Spatiotemporal appeared in Los Angeles, San
Feliciano, JT, between January 23- analysis Francisco, and Seattle.
Fodeh, SJ. March 24, 2020. • Social disruptiveness tweets were
most retweeted, and
implementation tweets were most
favourited.
Lai D, Wang 2020 USA Public Reddit 522 comments from an • Content analysis • “Symptoms” had highest number
D, Calvano J, Attitudes AMA session on of posts (27%), followed by
Raja AS, He coronavirus on March “Prevention” posts (25%)
S. 11, 2020, from 2:00 to • “Furthering discussion” was the
4:00pm EST. most common intent in
“Symptoms” posts (28%).
Li Y, Twersky 2020 USA Public Twitter 155,353 unique COVID- • Content analysis • Peril of COVID-19 was mentioned
S, Ignace K, Attitudes 19-related English tweets the most often, followed by mark,
Zhao M, posted from December responsibility, and group labeling
Purandare R, 31, 2019, to March 13, content.
Bennett-Jones 2020. • Tweets with conspiracy theories
B, Weaver were more likely included in group
SR. labeling and responsibility
information, but less in COVID-19
peril.
Lwin MO, Lu 2020 Singapore Public Twitter 20,325,929 tweets from • Sentiment analysis • Public emotions shifted strongly
J, Sheldenkar Attitudes 7,033,158 unique users from fear to anger over the course
A, Schulz PJ, from Jan 28 – April 9, of the pandemic, while sadness and
Shin W, 2020. joy also surfaced.
Gupta R, • Anger shifted from xenophobia at
Yang Y. the beginning of the pandemic to
discourse around the stay-at-home
notices.
• Sadness was highlighted by the
topics of losing friends and family
members, while topics related to
joy included words of gratitude
and good health.
• Emotion-driven collective issues
around shared public distress
experiences of the COVID-19
pandemic are developing and
include large-scale social isolation
and the loss of human lives.
Ma, R, Deng, 2020 China Public WeChat Top 200 accounts from • Simple linear • For nonmedical institution
Z, Wu, M. Attitudes 21 January 2020 to 27 regression accounts in the model, report and
January 2020. • Multiple linear story types of articles had positive
regression effects on users’ following
• Content analysis behaviors.
• For medical institution accounts,
report and science types had a
positive effect.
Medford RJ, 2020 USA Public Twitter 126,049 English tweets • Temporal analysis • The hourly number of COVID-19-
Saleh SN, Attitudes from 53,196 unique users • Sentiment analysis related tweets starkly increased
Sumarsono A, with matching hashtags from January 21, 2020 onward.
Perl TM, related to COVID-19 • Topic modeling • Fear was the most common
Lehmann CU. from January 14 to 28, using LDA emotion expressed in 49.5% of all
2020. tweets.
• The most common predominant
topic was the economic and
political impact.
Mohamad 2020 Brunei Public Twitter, 30 profiles from • Qualitative content • 5 narratives of local responses to
SM. Attitudes Instagram, individual profiles from analysis social distancing practices were
TikTok Instagram, Twitter, and apparent: the narrative of fear, the
TikTok. narrative of responsibility, the
narrative of annoyance, the
narrative of fun, and the narrative
of resistance.
Nguyen TT, 2020 USA Public Twitter 3,377,295 U.S. race- • SVM was used for • Proportion of negative tweets
Criss S, Attitudes related tweets from sentiment analysis referencing Asians increased by
Dwivedi P, November 2019 to June 68.4%.
Huang D, 2020. • Proportion of negative tweets
Keralis J, Hsu referencing other racial/ethnic
E, Phan L, minorities remained relatively
Nguyen LH, stable
Yardi I, • Common themes emerged during
Glymour MM, the content analysis of a random
Allen AM, subsample of 3,300 tweets include:
Chae DH, Gee racism and blame, anti-racism, and
GC, Nguyen, daily life impact
QC.
Odlum M, 2020 USA Public Twitter 2,558,474 Tweets from • Clustering • 15 topics (4 themes) were
Cho H, Attitudes January 21 to May 3, algorithm identified and visualized.
Broadwell P, 2020. • NLP • Positive sentiments, cohesively
Davis N, • Network diagrams encouraging online discussions,
Patrao M, and COVID-19 prevention
Schauer D, behaviors were uniquely observed
Bales ME, in African American Twitter
Alcantara C, communities.
Yoon S.
Park, HW, 2020 South Public Twitter 43,832 unique users and • Network analysis • Spread of information was faster in
Park, S, Korea Attitudes 78,233 relationships on • Content analysis the Coronavirus network than in
Chong, M. February 29, 2020. the other networks.
• Tweets containing medically
framed news articles were found to
be more popular than tweets that
included news articles adopting
nonmedical frames.
Qiang Chen, 2020 China Public Sina Weibo 1,411 posts pertinent to • Descriptive • Media richness negatively
Chen Min, Attitudes COVID-19 scraped from analysis predicted citizen participation via
Wei Zhang, "Healthy China," an • Hypothesis testing government social media, but
Ge Wang, official Sina Weibo dialogic loop facilitated
Xiaoyue Ma, account of the National engagement.
Richard Evans Health Commission of
China (NHCC) from
Jan 14 to Mar 5, 2020.
Samuel J, Ali 2020 USA Public Twitter 900,000 tweets from • Sentiment analysis • For short tweets, classification
N, Mokhlesur Attitudes February to March of packages, accuracy of 91% (NF) whereas
Rahman GG, 2020 • Textual analytics accuracy of 74% (logistic
Esawi, E, • ML classification regression).
Samuel, Y. methods: NF, • Both methods showed relatively
logistic regression weaker performance for longer
tweets.
Samuel, J, 2020 USA Public Twitter 293,597 tweets, 90 • Textual analytics to • For the reopening of the US
Mokhlesur Attitudes variables analyze public economy, there is more positive
Rahman GG, sentiment support sentiment support than negative
Ali N, Smaual • Sentiment analysis support.
Y, Pelaez A, using R package • Develops a novel sentiment polarity
Chong PHJ, Syuzhet based public sentiment scenarios
Yakubov M. (PSS) framework.
Su, Y, Xue, J, 2020 China, Italy Public Weibo, 850 Weibo users their • Wilcoxon tests • Individuals focused more on
Liu, X, Wu, P, Attitudes Twitter posts published from 9 “home” and expressed a higher
Chen, J, Chen, January 2020 to 5 level of cognitive process after a
C, Liu, T, February 2020. 14,269 lockdown in both Wuhan and
Gong, W, Tweets from 188 unique Lombardy.
Zhu, T. Twitter users from 23
 February 2020 to 21 • Level of stress decreased, and the
March 2020. attention to leisure increased in
Lombardy after the lockdown.
• Attention to group, religion, and
emotions became more prevalent
in Wuhan after the lockdown.
Thelwall, M., 2020 UK Public Twitter 3,038,026 English tweets • Word frequency • The results show that females are
Thelwall, S. Attitudes from Mar 10 – 23, 2020. comparison more likely to tweet about the
• Chi square analysis virus in the context of family,
social distancing and healthcare,
whereas males are more likely to
tweet about sports cancellations,
the global spread of the virus, and
political reactions.
Wang, T, Lu, 2020 China Public Weibo 999,978 randomly • Unsupervised • People concern 4 aspects regarding
K, Chow, KP, Attitudes selected COVID-19 BERT model: COVID-19:
Zhu, Q. related Weibo posts from classify sentiment a) The virus Origin
January 1, 2020 to categories. b) Symptom
February 18, 2020 • TF-IDF model: c) Production Activity
summarize the d) Public Health Control
topics of posts
• Trend analysis
• Thematic analysis
Wicke P, 2020 Ireland Public Twitter 203,756 tweets • Topic modeling • While the FAMILY frame covers a
Bolognesi Attitudes wider portion of the corpus, among
MM. the figurative frames WAR, a
highly conventional one, is the
frame used most frequently.
• Yet, this frame does not seem to be
apt to elaborate the discourse
around some aspects involved in
the current situation.
Xi W, Xu W, 2020 China Public Weibo 241 topics, their views • Thematic analysis • 6 themes were identified. The most
Zhang X, Attitudes and comments from • Temporal analysis prominent theme was
Ayalon L. January 20 to April 28, “Contributing to the community,”
2020 (n = 46, 24%) followed by “Older
 patients in hospitals” (n = 43,
23%).
• The theme “Contributing to the
community” was the most
dominant in the first phase
(January 20–February 20 as the
period of COVID-19 outbreak in
China).
• The “Older patients in hospitals”
theme was most dominant in the
second (February 21–March 17 as
the turnover period) and third
phase (March 18–April 28 as the
post-peak period in China).
Xie T, Tan T, 2020 China Public Baidu Number of cases by • Kendall’s Tau-B • Both the BAI and GT indices
Li J. Attitudes Attention February 29, 2020: rank test showed a similar trend in a slightly
Index 79,968 cumulative different way.
(BAI), confirmed cases; 41,675 • Daily GT was correlated to 7
Google cured cases; 2873 death indicators, whereas daily BAI was
Trends cases only correlated to 3 indicators.
(GT) • These indexes and rumors are
statistically related to disease-
related indicators.
• Information symmetry was also
found.
Xue J, Chen J, 2020 Canada Public Twitter 1,015,874 tweets from • LDA • 9 themes about family violence
Chen C, Hu R, Attitudes April 12 to July 16, 2020 were identified.
Zhu T.
Xuehua Han, 2020 China Public Sina Weibo 1,413,297 Weibo • Time series • Public response sensitive to the
Juanle Wang, Attitudes messages, including analysis epidemic and significant social
Min Zhang, 105,330 texts with • Kernel density events, especially in urban
Xiaojie Wang geographical location estimation agglomerations.
information, from 00:00 • Spearman
on Jan 9 to 24:00 on Feb correlation
10, 2020. • LDA model
• RF algorithm
Yigitcanlar T, 2020 Australia Public Twitter 96,666 tweets from • Descriptive • Social media analytics is an
Kankanamge Attitudes Australia in Jan 1 – May analysis efficient approach to capture
N, Preston A, 4, 2020. • Content analysis. attitudes and perceptions of the
Gill PS, • Sentiment analysis public during a pandemic.
Rezayee M, • Spatial analysis • Crowdsourced social media data
Ostadnia M, can guide interventions and
Xia B, decisions of the authorities during
Ioppolo G. a pandemic.
• Effective use of government social
media channels can help the public
to follow the introduced measures/
restrictions.
Yu J, Lu Y, 2020 Spain Public Twitter 22,223 tweets. • Topic modeling • Identifies 8 news frames for each
Muñoz- Attitudes • Network analysis newspaper’s Twitter account.
Justicia J. • The entire pandemic development
process is divided into 3 periods:
pre-crisis, lockdown, and recovery
period.

• Understand how Spanish news

media cover public health crises on
social media platforms
Zhao, Y., 2020 China Public Sina 4,056 topics from • Word segmentation • The trend of public attention can
Cheng, S., Yu, Attitudes Microblog December 31, 2019 to • Word frequency be divided into 3 stages.
X., Xu, H. hot search February 20, 2020. • Sentiment analysis • The hot topic keywords of public
list attention at each stage were
slightly different.
• The emotional tendency of the
public toward the COVID-19
epidemic-related hot topics
changed from negative to neutral,
with negative emotions weakening
and positive emotions increasing
as a whole.
• Divided the COVID-19 topics with
the most public concern into 5
categories: the situation of the new
cases of COVID-19 and its impact,
 frontline reporting of the epidemic
and the measures of prevention
and control, expert interpretation
and discussion on the source of
infection, medical services on the
frontline of the epidemic, and
focus on the worldwide epidemic
and the search for suspected cases.
Zhu B, Zheng 2020 China Public Weibo 1,858,288 microblog • LDA • “Double peaks” feature appears in
X, Liu H, Li J, Attitudes data the epidemic topic search curve.
Wang P. • The topic changes over time, the
fluctuation of topic discussion rate
gradually decreases.
• Political and economic centers
attract high attention on social
media.
• The existence of the subject of
“rumor” will enable people to have
more communication and
discussion
Table 2: Summary of chosen articles

Abbreviations:
* BERT = Bidirectional Encoder Representations from Transformers
* CDA: combination of depression and anxiety
* CDC: Centres for Disease Control and Prevention
* GQS = Global Quality Score
* JAMAS = Journal of the American Medical Association Score
* LDA: latent Dirichlet allocation
* MICI = Medical Information and Content Index
* MP: members of Parliament
* M-SDL: multiple-information susceptible-discussing-immune
* NB: naïve Bayes
* NLP: natural language processing
* RF: random forecast
* SME: social media exposure
* SMSI: social media search index
* SVM: support vector machine
* TCCI = Title–Content Consistency Index
* TF-IDF = Term Frequency-Inverse Document Frequency