Role of Social Media and Twitter
The COVID-19 pandemic caused widespread havoc worldwide as it spread rapidly and disrupted the regular social order. The pandemic necessitated extraordinary measures to contain its spread, resulting in restrictions on movement and interaction among people. As a result, a significant portion of the global population has resorted to using digital tools and platforms to conduct various activities.
At the same time, people were filled with curiosity and needed accurate, up-to-date information about the situation. They searched for and shared information online about the disease, and, in particular, social media became prevalent. Even with numerous social media platforms, Twitter has become an essential medium for sharing and obtaining information.
Governments and agencies found it to be an effective tool for making public new, verified data on infections, hospitalizations, and deaths. Top experts and researchers tweeted real-time knowledge and data, which policymakers and the public used to their benefit. Upon receiving this information, people also used Twitter to react and express their sentiments, mainly using emojis.
Role of Emojis
Emojis are digital symbols that convey simple yet eye-catching artwork, often featuring facial expressions. Emojis can be used to convey emotions among social media users who share a common understanding of them. Das (2020) notes that digital images facilitate non-verbal communication on social media, thereby enhancing the exchange of information and interaction among users.
Since their introduction around 2005, emojis have become popular tools for expressing joy, fear, celebration, and other emotions (Bai et al., 2019). Arora et al. (2020) state that emojis evoke different sentiments in digital communication through the messages people send via their devices and social media platforms, such as Twitter. Li et al. (2019) suggest that emojis are a distinct form of digital communication, possessing a unique social appeal. They have been developed out of an extra-linguistic origin, which is not considered part of the typical vocabulary of any ordinary semantic.
Therefore, emojis occupy a unique place in digital communication as they exemplify shared culture at a more granular level. For instance, the emoji that represents a smiley face is commonly used to convey friendliness or joy. Das (2020) notes that the general use of emojis varies depending on the discussion people are engaged in at a given time.
The prominence of emojis applies across many parts of the world, from the West to the East. According to Dyer and Kolic (2020), the state of California in the United States has been found to have a high concentration and variety of emoji use, particularly along its coast. The trend is associated with the high population density in the state’s central coastal cities, including San Diego, Los Angeles, Palo Alto, and San Francisco (Kejriwal et al., 2021). The high usage is also correlated with the area’s linguistic and cultural diversity, which promotes greater exchange and interaction.
Related Work
Empirical studies show that social media was a significant source of COVID-19-related data worldwide. Joang et al. (2019) refer to a cross-sectional research study that drew upon college students in Germany to assess their reliance on social media as a source of COVID-19-related information. They found that nearly 38 percent of learners frequently or occasionally turned to social media to seek information about the pandemic and related matters (Jiang et al., 2019).
Neely et al. (2021) conducted a similar research study, surveying 1,003 adults in the United States. Their findings revealed that slightly over three-quarters of respondents depended on social media to gather information about the COVID-19 pandemic. In addition, about 64 per cent of respondents considered the information accurate, suggesting that people should have taken the additional step of verifying it with a healthcare professional. Neely et al. (2021) also note that 59% of respondents visited social media at least once a week to scan through COVID-19-related information. The findings reveal that people trust social media as their primary source of information, despite concerns about the prevalence of fake news and misinformation.
Furthermore, research has shown that people can express their mental state on social media. Marzouki et al. (2022) analyzed several Facebook posts and found that people use the platform to express negative sentiments, such as depression, fear, anxiety, and stress. The majority of the participants typically portrayed these negative sentiments. According to Tran and Matsui (2022), people are increasingly using terms associated with negative feelings and irritation, primarily on Twitter. The findings also revealed that the participants manifested signs of depression as opposed to those who did not use social media.
Additionally, several other studies have shown that social media users tend to react adversely to the spread of COVID-19. Salvi et al. (2021) analyzed data from Twitter in Japan on social sensations toward COVID-19 for three months in 2020, from February to April 2020. They found that most people reported feeling a sense of fear regarding the infection.
Kaur et al. analyzed data from Twitter in February, May, and June 2020 (Shi et al., 2022). They established that most of the tweets portrayed negative sentiments. Dyer and Kolic (2020) also analyzed Twitter data and found proof of psychophysical distress. In particular, they found that users were more and more fixated on thinking of death, and there was reduced emotional expression. The research shows that social media, including Twitter, can cause mental illness due to heightened anxiety, stress, and infectious panic (Tran & Matsui, 2022). The situation was worsened by misinformation on COVID-19 and fake news that tended to overstate the perceived threat.
Findings on people’s reactions through social media suggest that a comparable long-term trend may exist across different regions. Tran and Matsui (2022) compared social media reactions across six countries: South Korea, Indonesia, India, Germany, and Thailand. They found a similar pattern based on social media reactions. The similarity in the findings was further supported by the consideration that the countries have a high Twitter usage.
Yamashita and Yokoyama (2022) also note that tweets can be easily gathered using the primary languages spoken by citizens. The observed trend revealed a dramatic increase in Twitter usage during the COVID-19 pandemic (Suntwal et al., 2022). However, a significant decline was observed afterward, even as each country encountered fresh waves at various times, leading to a surge in reactions on social media. Nevertheless, the general observation was a decline in social media reactions in 2022 across all countries (Kaur et al., 2022). Therefore, the general trend was that social media reaction to the COVID-19 pandemic was exceptionally high at the onset and slowly faded away over time.
Method
The data for this research study emanates from social media, particularly Twitter, in Japan. The platform is the preferred source because of its popularity in the country, promotes a high level of public interaction, and can easily attract participation from unfamiliar people. According to Tran and Matsui (2022), the 2022 data indicated that Japan has the second-highest number of Twitter users, after the United States, with more than 50 million users. The platform also offers a more accessible option for collecting and analyzing historical data, as the company provides an API for research work. The data considered for the research comprised COVID-19-related tweets and sentimental expressions through emojis.
Sentiment analysis was part of the study to understand the trends in people’s emotional expressions on Twitter following the COVID-19 pandemic. It involves using an algorithm to define people’s views on a specific topic. The nature of this research study is particularly relevant given the rise of social media usage that has drawn the attention of many organizations seeking to gain a deeper understanding, particularly to monitor and manage their outlook and presence (Kejriwal, 2021).
Sentiment evaluation is a crucial undertaking in Natural Language Processing (NLP), one of the key areas in Artificial Intelligence (Yamashita & Yokoyama, 2022). It facilitates the understanding, interpretation, and utilization of human languages by computers. This is realized through the conversion of unstructured text into organized data. The study will apply Bidirectional Encoder Representations from Transformers (BERT) to build a sentiment classifier (Tran & Matsui, 2022). The model developed by the Google AI Language team in 2018 is bidirectional and contextual and is considered a significant breakthrough in the field.
Results
The data indicate that Japanese people commonly use facial expressions and emojis in their tweets. They mainly involve expressions of fear, sadness, happiness, and anger. Their emotional manifestation is connected to the sentimental behavioral changes brought about by the spread of the COVID-19 pandemic. Therefore, social media captured their accurate and compelling emotional reactions based on data from the most popular emojis.
Emojis were shown to provide vital sentiment indicators, which are often missing or indefinite, in the text alone. Certain emojis were identified as consistently correlating with specific sentiments. The analysis also revealed distinct regional trends in emoji usage, reflecting Japan’s varied emotional landscape across its regions. However, with the containment of COVID-19, there was a decline in social media usage in emotional expression connected to the disease.
Conclusion
The findings indicate that emojis can be utilized to analyze public sentiment on social media regarding a specific topic. This suggests the possibility of monitoring public emotion for any undertaking. However, more work needs to be done in describing the various meanings of emojis. This makes it highly challenging to correctly apply and analyze to obtain an accurate sentiment analysis of social media users, particularly regarding COVID-19 and in general. Emojis on social media still provide a critical basis for understanding and predicting people’s emotions during a pandemic. Even as the use of emojis to express sentiments related to the COVID-19 pandemic declines, there is a need for further research to establish their role in facilitating communication and interaction among people.
References
Arora, A., Chakraborty, P., Bhatia, M. P. S., & Mittal, P. (2020). Role of emotion in excessive use of Twitter during COVID-19 imposed lockdown in India. Journal of Technology in Behavioral Science, 6, 370-377.
Bai, Q., Dan, Q., Mu, Z., & Yang, M. (2019). A systematic review of emoji: Current research and future perspectives. Frontiers in Psychology, 10.
Das, A. (2020). How has the coronavirus (COVID-19) pandemic affected global emoji usage? Journal of Human Behavior in the Social Environment, 31, 425 – 434.
Dyer, J., & Kolic, B. (2020). Public risk perception and emotion on Twitter during the Covid-19 pandemic. Applied Network Science, 5(1).
Jiang, J., Thomason, J., Barbieri, F., & Ferrara, E. (2022). Geolocated social media posts are happier: Understanding the characteristics of check-in posts on Twitter.
Kaur, S., Kaul, P., & Zadeh, P. M. (2020). Monitoring the dynamics of emotions during COVID-19 using Twitter data. Procedia Computer Science, 177, 423–430.
Kejriwal, M., Wang, Q., Li, H., & Wang, L. (2021). An empirical study of emoji usage on Twitter in linguistic and national contexts. Online Social Networks and Media, 24, 100149.
Li, M., Chng, E., Chong, A. Y. L., & See, S. (2019). An empirical analysis of emoji usage on Twitter. Industrial Management & Data Systems, 119(8), 1748–1763.
Marzouki, Y., Aldossari, F. S., & Veltri, G. A. (2021). Understanding the buffering effect of social media use on anxiety during the COVID-19 pandemic lockdown. Humanities and Social Sciences Communications, 8(1).
Neely, S., Eldredge, C., & Sanders, R. (2021). Health information seeking behaviors on social media during the COVID-19 pandemic among American social networking site users: Survey Study. Journal of Medical Internet Research, 23(6), e29802.
Salvi, C., Iannello, P., Cancer, A., McClay, M., Rago, S., Dunsmoor, J. E., & Antonietti, A. (2021). Going viral: How fear, socio-cognitive polarization and problem-solving influence fake news detection and proliferation during COVID-19 pandemic. Frontiers in Communication, 5.
Shi, W., Zeng, F., Zhang, A., Tong, C., Shen, X., Liu, Z., & Shi, Z. (2022). Online public opinion during the first epidemic wave of COVID-19 in China based on Weibo data. Humanities and Social Sciences Communications, 9(1).
Suntwal, A., Brown, S. J., & Brandimarte, L. (2021). Pictographs, Ideograms, and Emojis (PIE): A Framework for Empirical Research Using Non-verbal Cues.
Tran, V., & Matsui, T. (2022). Tweet analysis for enhancement of COVID-19 epidemic simulation: A case study in Japan. Frontiers in Public Health, 10.
Yamashita, O., & Yokoyama, S. (2022). Preference aware route recommendation using one billion geotagged tweets.