Research Paper

How_did_Twitter_Users_React_to_the_Pande.pdf

Home >Information Systems homework help >Research Paper

Asia Pacific Journal of Multidisciplinary Research, Volume 8, No. 3, August 2020 _________________________________________________________________________________________________________________________

52 P-ISSN 2350-7756 | E-ISSN 2350-8442 | www.apjmr.com

How did Twitter Users React to the

Pandemic? Social Network Analysis of Public

Tweets on CoViD-19 Outbreak

Jasten Keneth D. Treceñe 1 , Ralph Jerico P. Abides

1 Eastern Visayas State University – Tanauan Campus,

2 Visayas State

University – Tolosa Campus, Philippines

[email protected] 1 , [email protected]

Date Received: May 7, 2020; Date Revised: July 15, 2020

Asia Pacific Journal of

Multidisciplinary Research

Vol. 8 No.3, 52-59

August 2020

P-ISSN 2350-7756

E-ISSN 2350-8442

www.apjmr.com

ASEAN Citation Index

Abstract –Several researchers have presented several studies on the CoViD-19 outbreak like on the

epidemiological aspects of the disease, diagnostics method of the novel coronavirus, clinical characteristics,

transmission, and vaccines. However, the sentiments and behaviour of the people online particularly in

twitter remain unexplored. In this paper we focused on exploring peoples’ tweets to uncover their attitudes,

sentiments, and find out the network effects of peoples’ tweets and the heated topics.Text mining approach

was utilized using sentiment and social network analysis. Term document matrix, word cloud, nrc_sentiment

dictionary, histogram, community edge betweenness algorithm, and network graph were used in the study.

An API account was created wherein15000 tweets were extracted from March 22, 2020 to March 31, 2020

containing the keyword #COVID-19 to make a working data for analysis. Results from the social network

analysis showed a close relationship between tweets where people are globally talking part by sharing

information about the CoViD-19. The peoples’ attitude showed the willingness to follow government

precautionary measures to lessen the impact of the virus. Despite of the fear and sadness felt by the people

over twitter, sentiment analysis revealed positive emotion towards the crisis. Such insights are significant

when guiding people to respond appropriately and helping them to learn to cope with the sudden infectious

disease as it promotes social stability. This will also help the authorities understand the sentiments and

anxieties of the people, giving a strong direction to enact policies beneficial to the people. Moreover, social

network analysis can be used as a method of understanding the behaviour of the people online and how these

people are talking towards an issue.

Keywords –Social Network, Sentiment Analysis, Text Mining, covid-19, coronavirus

INTRODUCTION

As the new coronavirus outbreak hits the world and

people need to stay at home to avoid of being infected

with the virus and lessen the spreading of the contagion,

thus where most of the conversations are taking place

online. People take the opportunity of using internet to

share information, raise their concerns, and consume

most of their time in the internet while in quarantine.

The time when those online discussions light up also

tell us a lot about how their feelings around the

pandemic are growing. With the advent and the rapid

growth of technology, there has been a considerable

change on the information landscape and information-

consumption of the people [1]. Discussion of the

CoViD-19 has been flooded across various social media

platforms as reported by media analytics [2].

At the early stage, literatures emphasized that we

still have limited data about the outbreak, this can be

found in the study of Fong where they only have a

handful of datasets to develop a model. This is also

because we still have few studies about the disease [24].

It is important to know the peoples’ sentiments to

CoViD-19 during the current situation. Such insights

are significant when guiding people to respond

appropriately and helping them to learn to cope with the

sudden infectious disease as it will also promote social

stability. Furthermore, this study will also help

authorities to know peoples’ worries and anxieties,

having them a strong direction and ratify new policies

helpful to the people. This study used various text

mining techniques and algorithms mainly, sentiment

analysis using the nrc_sentiment dictionary in R and

network analysis using community edge betweenness.

The rise of the discussions of the corona virus online

has been followed as the pandemic has been infecting

more and more people around the world. Sprinklr, a

media analytics noted that several emotions were also

been expressed online based on the emojis that most

Treceñe et al., How did Twitter Users React to the Pandemic? Social Network Analysis of Public Tweets… _________________________________________________________________________________________________________________________

53 P-ISSN 2350-7756 | E-ISSN 2350-8442 | www.apjmr.com

Asia Pacific Journal of Multidisciplinary Research, Vol. 8, No. 3, August 2020

commonly associated with the corona virus. Peoples

concern on the corona virus is evidently seen in the

search history of Google. An exponential increase of

search terms relating to corona virus has been recorded.

With the different analytics presented by Sprinklr and

google trends, twitter has been also widely used by

many people and as basis also for understanding the

most talked issue around the world. Twitter has

developed rapidly in recent years, increasing number of

public individuals are already using this social media

platform to communicate, share information, and raise

their concerns and opinions towards a specific issue.

Twitter has become an important channel for promoting

risk communication during crisis [3-4]. The use of

social media particularly twitter to measure public

attention has also been gradually applied in research on

infectious diseases [3]-[7], [9].

Currently, the world is experiencing a Corona virus

(CoViD-19) outbreak and has now spread to more than

50 countries [10]. It was already declared by WHO as a

pandemic and a Public Health Emergency of

International Concern (PHEIC). With the onset of

CoViD-19, many people are turning to twitter to assess

the severity of the situation, raise concerns to the

current condition, and to the government policies and

actions. Presently, various text mining techniques

particularly sentiment and social network analysis has

become an important tool for understanding people’s

behaviour online and come up with meaningful insights

from them [3]-[4]-[5]. Various researchers made efforts

in different aspects to fight against CoViD-19 and

promote the prevention and mitigation of the pandemic

like on the epidemiological aspects of the disease,

diagnostics method of the novel corona virus, clinical

characteristics of the disease, disease transmission, and

virus vaccines [11]-[13]. A study on the CoViD-19

outbreak submitted to the bulletin of World Health

Organization used the predictive modelling approach to

forecast CoViD-19 outbreak within and outside China

based on daily observation [14]. They also analysed the

sentiments from news articles and classify these articles

based on the polarity, this is also to understand the

influence of the news to the behaviour of the people,

politically and economically.

Pastor [15] also studied on the sentiments to the

CoViD-19 pandemic. Both qualitative and quantitative

method was used in the study with the application of

sentiment analysis. However, the study was just limited

to only a specific group of people, similar to the

research who also studied in the CoViD-19 outbreak

where they analysed the sentiments of Chinese from the

extracted data in a microblog hot search list [5]. But the

study focused only on a microblog wherein more people

are accessing other social media platform like twitter. In

this paper we focused on peoples’ tweets around the

world from March 22, 2020 to March 31, 2020 wherein

people express more of their opinions in the site.

The rest of the paper is structured as follows:

Section II presents the objectives of the study, section

III elucidates the methodology of the study, including

the research approach used, the data collection, the

research process and the data analysis, section IV

illustrates the results of the study, it also includes the

data exploration, sentiment analysis, and network

analysis, section V provides the summary of findings

and the discussions, section VI summarizes the

conclusions, and finally section VII presents the

limitations of the study and future works..

OBJECTIVES OF THE STUDY

The propagation of social media usage for

discussion of opinions and feelings by the public has

created possibilities of analyzing such sentiments about

any prevalent discourse. This study analyzed the

sentiments and attitudes about the CoViD-19 pandemic

expressed globally over twitter. Specifically, this study

explored the data towards peoples’ attitude on the

CoViD-19 pandemic, analyzed and presented the

sentiments of people towards the pandemic, and

identified how people’s tweets are closely connected to

each other using community detection algorithm,

identified the most influential words inside the graph

using the measures of betweenness centrality and

degree, and find out the heated topics.

MATERIALS AND METHODS

Research Approach

To achieve the objectives of the study, the

researchers used the text mining techniques such as

sentiment analysis (SA) and network analysis. The text

mining area has been widely used in computer science

which adopts the concepts of natural language

processing, knowledge management, data mining, and

machine learning [16]. It explores interesting patterns

from the useful unstructured data that has been

extracted [17].

Data Collection

The tweets were extracted into a working data for

analysis using R programming. Also, an API account on

tweeter was created first to allow us to harvest tweets.

We extracted 15000 tweets from different tweeter users

54 P-ISSN 2350-7756 | E-ISSN 2350-8442 | www.apjmr.com

Asia Pacific Journal of Multidisciplinary Research, Vol. 8, No. 3, August 2020

globally from March 22, 2020 to March 31, 2020

containing the keyword #COVID-19for the website.

Replies and retweets were not included. The data

consists of 16 columns and 15000 rows where it

contains the tweet, followed by information such as the

like engagement, time, user id, and etc.

Research Process and Data Analysis

The research process includes of extracting first the

tweets, after, the data will undergo on the pre-

processing and data cleaning stage, then will go on the

process of analysing the sentiments and network

analysis together with data visualization.

Fig. 2. Research process

Data Pre-processing and Cleaning

The data set is transformed into a corpus, a corpus

is a group of text known in R. Then, the corpus, was

pre-processed using tokenization and text

normalization. This stage is very important when

dealing with large amount of data.

Tokenization–In this stage of pre-processing, all

the characters were transformed into lowercase,

punctuations and numbers were removed, English stop

words and white spaces were also removed. Moreover,

uniform resource locator, emojis, and unnecessary

words were also removed such as names mentioned in

tweets.

Text normalization – before further processing of

the text, it needs to be normalized. It is generally

referring to allowing the words on equal footing and

allows the processing to continue uniformly. Two tasks

were used to normalize the text such as text stemming

and lemmatization. In the stemming process, words like

need, needed, and needing were stemmed to the word

“need”. In the lemmatization part, words like corona

virus, ncov and virus were transformed into its citation

form to “covid”. This idea is used to reduce the distinct

number of words in the corpus that will improve the

analysis.

Sentiment Analysis

After the pre-processing stage, sentiment analysis

was done to reveal the emotions behind people’s tweets.

Sentiment analysis (SA) is a natural language process

that creates meaningful information out of the textual

data [18]. The technique was used to identify the

emotions expressed by the people from the tweets.

These emotions focused on eight emotions such as trust,

joy, sadness, fear, anger, surprise, disgust and

anticipation[18]. To obtain the sentiment scores of the

tweets, “nrc_sentiment” dictionary was used to

calculate the presence of eight emotions and their

corresponding valence. The sentiment analysis helped

to learn individual’s emotion and attitudes to the

CoViD-19 outbreak.

Network Analysis

In text network analysis, a text is represented as

graph. It helps identifies relationships of text in social

media platforms. The words are the nodes and co-

occurrences of the words are the connections between

them [19]. Then, the community detection algorithm

was used on the constructed graph to identify the groups

of nodes that are more densely connected to one another

than to the rest of the network as well as the most

influential words inside the graph using the measures of

betweenness centrality and degree.

To avoid the messy display of the data, we only

cover terms that appeared more than 30 times in the

text. The connected terms are those that appear together

on Twitter. Then edge betweenness was utilized to

cluster all the words. Betweenness signifies how

recurrently a node is between other nodes’ paths. Edge

betweenness is the number of shortest paths that go

through an edge in a network graph [20].

Data Visualization

Different data visualization techniques were used

to obtain the objectives of the study. First, to know the

attitude of the people to the CoViD-19 pandemic, the

Extracting

Tweets

Data pre-processing

and Cleaning

Sentiment

Analysis Network

Analysis

Data Visualization

55 P-ISSN 2350-7756 | E-ISSN 2350-8442 | www.apjmr.com

Asia Pacific Journal of Multidisciplinary Research, Vol. 8, No. 3, August 2020

term document matrix (TDM) was used and presented

in a word cloud. Second, to identify the sentiment of the

people based on their tweets, the sentiment analysis

technique was used in R environment. The

“nrc_sentiment” dictionary was used to obtain the

sentiment scores and their valence. Lastly, to visualize

the network of tweets, histogram and tweet vertices

were used to illustrate how tweets are closely connected

to each other. Finally, term visualization, community

edge betweenness and network graph was used to

identify heated topics within the extracted tweets.

Ethical Considerations

As twitter becomes a popular social networking

site where it offers free advanced programming

interface that allows access to millions of tweets,

including the metadata on the user’s exact physical

location, a careful data handling practices have been

applied. The objectives and methodologies were

discussed clearly, the anonymity of tweet authors

remain protected, and personal and private twitter data

were omitted.

RESULTS

Data Exploration

In the following section, the researchers presented

the results obtained from the 15000 extracted tweets

using R programming analysed in R Studio

environment.

Fig. 3. Frequency of terms

Peoples’ tweets focused on the cases of the CoViD-

19.They talked about new positive cases of CoVid-19

and the exponential increase in just a short period. They

also talked of being prepared for a higher number of

CoViD-19 infected cases. Figure 3 shows the words like

“tested”, “help”, “need”, “best”, and “said”, where it

gives us an indication of their attitude towards the

disease. From the extracted tweets, people are taking

part by sharing information about the virus. They call to

help each other, give some prayers and help the

government by following the government’s

precautionary measures. As seen in figure 3, the most

frequent terms in the corpus are “cases”, “positive”, and

“test”. Sample tweets showed discussions on people

who were tested positive of the virus.

Fig. 4. Word cloud of terms

The word cloud displays the frequent terms

mentioned in the tweets. The terms shown in the word

cloud was based on the generated TDM using R

programming. Cases as the most mentioned, followed

by positive and test, need, cough and as also indicated

in figure 3.

Sentiment Analysis of People’s Tweets

To obtain the sentiment scores and the valence,

“nrc_sentiment” dictionary in R was used. It helps

captured the people’s emotion in the corpus.

Fig. 5. Sentiment score of CoViD-19 tweets

Figure 5 are the emotions expressed by the public

in tweets. As shown in the figure, trust has the highest

sentiment score, followed by anticipation, fear, sadness,

joy, and anger, while disgust and surprise have the least

sentiment scores. However, the valence of emotions

expressed by the people from the tweets remained

positive as shown in the figure.

56 P-ISSN 2350-7756 | E-ISSN 2350-8442 | www.apjmr.com

Asia Pacific Journal of Multidisciplinary Research, Vol. 8, No. 3, August 2020

Network Analysis of Tweets

Fig. 6. Histogram of degree nodes

The histogram in figure 6 shows the degree values

of the tweets. It indicates that the histogram is skewed

on the right for most degree of values of the tweets.

Few extreme values are seen in the other side of the

histogram. This implies that terms in the tweets have

close connection with others.

Fig. 7. Term visualization

The network graph in figure 7 provides the cleaner

look of the terms using only the terms with the

frequency of more than 30. The connected terms above

are those terms appeared together on the tweets. The

term health is probably at the centre of the network

graph and closely connected to other terms such as

home, patient, months, case, and number.

Fig. 8. Nodes clustering based on edge betweenness

Figure 8 indicates the clustered networked terms

based on edge betweenness. The four clusters talk about

the various measures to lessen the impact of the

coronavirus, CoViD-19 cases across countries, the need

to be tested and understand the disease, the need of

protective equipment, and the common symptoms of the

disease.

Fig. 9. Tweets vertices

The plot in figure 9 shows the network impact of

tweets and its distribution. The points that are already

far from the dense area of the plot shows no connection

among others while points near and at the centre are

tweets that are related. It shows that only few from the

extracted tweets do not have connection to each other.

Fig. 10. Network of tweets

The network graoh in figure 10 shows the

detailed network of tweets. The numbers in each of the

points show the twitter ID of tweets in raw data. Tweets

in the dense area are most recurrently liked, retweeted,

and commented. Below are the tweets randomly picked

in every dense area to see what people are commonly

57 P-ISSN 2350-7756 | E-ISSN 2350-8442 | www.apjmr.com

Asia Pacific Journal of Multidisciplinary Research, Vol. 8, No. 3, August 2020

talking about the CoViD-19 outbreak. Based on the

most talked tweets online, we now understand how

people react to the current situation the world is facing.

People are talking what are the best way to tell

if someone has been infected by the disease. People are

taking part on sharing information about CoViD-19

confirmation tests. Giving awareness to others by

sharing information of the current number of CoViD-19

cases. They believe that the measures and actions made

by the government will help lessen the impact of the

disease. Reminding people to be prepared as the cases

of infected people are increasing rapidly. Lastly, they

are talking about the effects of the community

quarantine.

DISCUSSIONS

People’s Attitudes to the CoViD-19 Pandemic

People are taking part on this crisis by sharing

information on twitter, however, subject to this is the

proliferation of false news. With the spread of the

disease is also the spread of false news. Several social

media platforms are making their moves to live up their

responsibilities as they have the medium of what

information should appear on their sites. Twitters are

doing their ways to lessen the spread of false news [21].

Moreover, people should be mindful of sharing

information in social media particularly twitter, as the

study on the sentiment of tweets on covid-19 confirms

that there are misleading stories tend to misinform

readers [22]. They are also actively talking about how

different private individuals, businesses and

governments are doing to help lessen the impact of the

disease. Lastly, they call everyone the need for prayers,

help each other and follow the precautionary measures

imposed by the authorities because of the rapid spread

of the disease.

People’s Sentiments towards CoViD-19 Pandemic

They trust on the measures imposed by the

government that it will make us free from the disease.

They also express fear of being infected especially

those who are more vulnerable like the children and old

ones. Moreover, fear was also expressed by the people

for the front-liners of being infected by the disease

especially the doctors and nurses. Fear was also an

expressed emotion because of the rapid spread of the

disease and their still no clear treatment and vaccine for

the corona virus [26]. However, this is on contrary to an

article, which she claims that people are becoming less

fearful. Based on the analysis of tweets, people are not

anymore expressing fear about the corona virus, they

become more knowledgeable about the disease [23].

This is also in support of the attitudes revealed by the

people, this is because of constantly sharing of

information about the contagion. Nevertheless, subject

to this is a confirmatory study about the emotions of

people expressed based on analysis of tweets. This

study revealed an interesting result where despite of the

crisis, an optimistic emotion was more expressed from

the outcome of the analysis. This is similar to a study in

India where the results of the sentiment analysis

revealed a positive emotion toward the covid19

outbreak [25][27]. They also trust their government

that the measures implemented will be successful and

people will not struggle [25].

Heated Topics based on the Network of Tweets

Based on the results of the histogram and tweet

vertices, it showed that most of the tweets have close

connection with each other. The tweets were clustered

into four; it revolves on the discussions about

precautionary measures to lessen the impact of the

corona virus, cases who are infected by the disease, the

need to be tested and understand the disease, and the

need to have enough protective equipment.

CONCLUSION AND RECOMMENDATION

Text mining has been widely used across fields,

from business, education, and in health issues. This

study used the approach particularly sentiment and

social network analysis to uncover the attitudes and

sentiments of the people towards the CoViD-19

pandemic. This study also looked into how the tweets

are connected to each other and find out the popular

topics. Results of the study showed how people are

taking part on the crisis, by sharing reliable information

for the awareness of everyone, calling to help each

other and follow the precautionary measures imposed

by the authorities. Interestingly, despite the fear felt by

the people, the sentiment scores revealed positive

emotion towards the crisis. As twitter has been widely

used by researchers in various fields, knowing the

sentiments and what people are talking online will help

authorities understand what is happening and how is

people reacting to the current situation, this will also

help enact policies that would be valuable to the

everyone.

LIMITATIONS AND FUTURE WORKS

Based on the results of the study, the different text

mining approach used successfully revealed the

attitudes and sentiments of the people toward the

58 P-ISSN 2350-7756 | E-ISSN 2350-8442 | www.apjmr.com

Asia Pacific Journal of Multidisciplinary Research, Vol. 8, No. 3, August 2020

CoViD-19 pandemic. It was also successfully identified

the network of tweets and the popular topics. However,

there are improving points of the study. First, the data

are limited only to the number of tweets. Second, the

cluster analysis used only one algorithm to group the

terms. Future works could extract a larger number of

tweets to gain more interesting results of the study.

They can also consider comparing other algorithms for

cluster analysis.

REFERENCES [1] Kwak, H., Lee, C., Park, H., & Moon, S. (2010, April).

What is Twitter, a social network or a news media?. In

Proceedings of the 19th international conference on

World wide web (pp. 591-

600).https://doi.org/10.1145/1772690.1772751

[2] Molla R. (2020), “How coronavirus took over social

media”, Voxmedia. Accessed March 20, 2020 [online].

Available at https://www.shorturl.at/qDJT1

[3] Househ, M. (2016). Communicating Ebola through

social media and electronic news media outlets: A

cross-sectional study. Health informatics journal, 22(3),

470-478.doi: 10.1177/1460458214568037

[4] Gui, X., Wang, Y., Kou, Y., Reynolds, T. L., Chen, Y.,

Mei, Q., & Zheng, K. (2017). Understanding the

Patterns of Health Information Dissemination on Social

Media during the Zika Outbreak. In AMIA Annual

Symposium Proceedings (Vol. 2017, p. 820). American

Medical Informatics Association.Available at

https://www.ncbi.nlm.nih.gov/pubmed/29854148

[5] Zhao, Y., & Xu, H. (2020). Chinese public attention to

COVID-19 epidemic: Based on social media.

medRxiv.doi:

https://doi.org/10.1101/2020.03.18.20038026

[6] Wong, R., & Harris, J. K. (2018). Geospatial

distribution of local health department tweets and online

searches about Ebola during the 2014 Ebola outbreak.

Disaster medicine and public health preparedness,

12(3), 287-290.doi: 10.1017/dmp.2017.69.

[7] Seltzer, E. K., Horst-Martz, E., Lu, M., & Merchant, R.

M. (2017). Public sentiment and discourse about Zika

virus on Instagram. Public Health, 150, 170-175.doi:

10.1016/j.puhe.2017.07.015.

[8] Fung, I. C. H., Fu, K. W., Chan, C. H., Chan, B. S. B.,

Cheung, C. N., Abraham, T., & Tse, Z. T. H. (2016).

Social media's initial reaction to information and

misinformation on Ebola, August 2014: facts and

rumors. Public Health Reports, 131(3), 461-473.doi:

10.1177/003335491613100312

[9] Fu, K. W., Liang, H., Saroha, N., Tse, Z. T. H., Ip, P., &

Fung, I. C. H. (2016). How people react to Zika virus

outbreaks on Twitter? A computational content

analysis. American journal of infection control, 44(12),

1700-1702.doi: 10.1016/j.ajic.2016.04.253

[10] WHO, "Advice for Public," WHO Int., 2020. [Online].

Available:https://www.ibit.ly/08HM [Accessed 27

February 2020].

[11] World Health Organization, "Report of the WHO-China

Joint Mission on Coronavirus Disease 2019 (COVID-

19)," World Health Organization, 2020. Available

athttps://www.is.gd/nFMe79

[12] Lai, C. C., Shih, T. P., Ko, W. C., Tang, H. J., & Hsueh,

P. R. (2020). Severe acute respiratory syndrome

coronavirus 2 (SARS-CoV-2) and corona virus disease-

2019 (COVID-19): the epidemic and the challenges.

International journal of antimicrobial agents,

105924.doi:

https://doi.org/10.1016/j.ijantimicag.2020.105924

[13] Kementerian Kesihatan Malaysia, "COVID-19,"

Kementerian Kesihatan Malaysia, 2020. [Online].

Available:

https://www.facebook.com/kementeriankesihatanmalay

sia/.[Accessed 27 February 2020].

[14] Hamzah F. A. B., Lau C. H., Nazhri H., Ligot D. V., G.

Lee, C. L. Tan, M. K. B. Mohd Shaib, U. H. B. Zaidon,

A. B. Abdullah, M. H. Chung, C. H. Ong, P. Y. Chew,

R. E. Salunga (2020). CoronaTracker: Worldwide

COVID-19 Outbreak Data Analysis and Prediction.

[Submitted]. Bull World Health Organ. doi:

http://dx.doi.org/10.2471/BLT.20.255695

[15] Pastor, C. K. L. (2020). Sentiment Analysis on

Synchronous Online Delivery of Instruction due to

Extreme Community Quarantine in the Philippines

caused by COVID-19 Pandemic. Asian Journal of

Multidisciplinary Studies, 3(1), 1-

6.https://asianjournal.org/online/index.php/ajms/article/

download/207/89

[16] Radovanović M., & Ivanović M.” Text mining:

Approaches and applications”. Novi Sad J. Math, 2008,

38(3), 227-234. Available at ibit.ly/O9Z1

[17] Feldman, R., & Sanger, J. (2007). The text mining

handbook: advanced approaches in analyzing

unstructured data. Cambridge university press.Available

at ibit.ly/zKEI

[18] Treceñe, J. K. (2019). Delving The Sentiments To

Track Emotions In Gender Issues: A Plutchik-Based

Sentiment Analysis In Students' Learning

Diaries.International Journal of Scientific &

TechnologyResearch, 8(12), 1134 – 1139. doi:

10.13140/RG.2.2.35883.80164.

[19] Paranyushkin, D. (2019, May). Infranodus: generating

insight using text network analysis. In The World Wide

Web Conference (pp. 3584-3589). Doi:

https://doi.org/10.1145/3308558.3314123

[20] Girvan M, & Newman ME (2002). Community

structure in social and biological networks. Proceedings

of the national academy of sciences. 99(12):7821-6.doi:

https://doi.org/10.1073/pnas.122653799

59 P-ISSN 2350-7756 | E-ISSN 2350-8442 | www.apjmr.com

Asia Pacific Journal of Multidisciplinary Research, Vol. 8, No. 3, August 2020

[21] Coronavirus Fake News: How Facebook, Twitter, And

Instagram Are Tackling The Problem. http://ibit.ly/yi1a,

Date accessed March 31, 2020

[22] How did Twitter react to the Coronavirus pandemic?.

http://ibit.ly/nEwa, Date accessed March 31, 2020

[23] David, T. (2020), “What social media tells us about

sentiment during COVID-19”, PR Week Region Asia.

Accessed March 31, 2020 [online]. Available at

http://ibit.ly/9rbA

[24] Fong, S. J., Li, G., Dey, N., Crespo, R. G., & Herrera-

Viedma, E. (2020). Finding an accurate early

forecasting model from small dataset: A case of 2019-

ncov novel coronavirus outbreak. arXiv preprint

arXiv:2003.10776.

[25] Barkur, G., & Vibha, G. B. K. (2020). Sentiment

Analysis of Nationwide Lockdown due to COVID 19

Outbreak: Evidence from India. Asian Journal of

Psychiatry. https://doi.org/10.1016/j.ajp.2020.102089

[26] Abd-Alrazaq, A., Alhuwail, D., Househ, M., Hamdi,

M., & Shah, Z. (2020). Top Concerns of Tweeters

During the COVID-19 Pandemic: Infoveillance Study.

Journal of Medical Internet Research, 22(4),

e19016.http://www.jmir.org/2020/4/e19016/doi:

10.2196/19016

[27] Dubey, A. D. (2020). Twitter Sentiment Analysis

during COVID19 Outbreak. Available at SSRN

3572023.Available at SSRN:

https://ssrn.com/abstract=3572023 or

http://dx.doi.org/10.2139/ssrn.3572023

COPYRIGHTS

first publication rights granted to APJMR. This is an open-

access article distributed under the terms and conditions of

the Creative Commons Attribution license (http://creative

commons.org/licenses/by/4).