This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Infodemiology, is properly cited. The complete bibliographic information, a link to the original publication on https://infodemiology.jmir.org/, as well as this copyright and license information must be included.
A global rollout of vaccinations is currently underway to mitigate and protect people from the COVID-19 pandemic. Several individuals have been using social media platforms such as Twitter as an outlet to express their feelings, concerns, and opinions about COVID-19 vaccines and vaccination programs. This study examined COVID-19 vaccine–related tweets from January 1, 2020, to April 30, 2021, to uncover the topics, themes, and variations in sentiments of public Twitter users.
The aim of this study was to examine key themes and topics from COVID-19 vaccine–related English tweets posted by individuals, and to explore the trends and variations in public opinions and sentiments.
We gathered and assessed a corpus of 2.94 million COVID-19 vaccine–related tweets made by 1.2 million individuals. We used CoreX topic modeling to explore the themes and topics underlying the tweets, and used VADER sentiment analysis to compute sentiment scores and examine weekly trends. We also performed qualitative content analysis of the top three topics pertaining to COVID-19 vaccination.
Topic modeling yielded 16 topics that were grouped into 6 broader themes underlying the COVID-19 vaccination tweets. The most tweeted topic about COVID-19 vaccination was related to vaccination policy, specifically whether vaccines needed to be mandated or optional (13.94%), followed by vaccine hesitancy (12.63%) and postvaccination symptoms and effects (10.44%) Average compound sentiment scores were negative throughout the 16 weeks for the topics
Identification of dominant themes, topics, sentiments, and changing trends about COVID-19 vaccination can aid governments and health care agencies to frame appropriate vaccination programs, policies, and rollouts.
Since the outbreak of COVID-19, caused by the SARS-CoV-2 virus, in November 2019, the pandemic continues to pose a serious threat to the lives of millions of individuals around the globe. By June 2021, the virus had infected over 176 million individuals, resulting in over 3.8 million deaths worldwide [
Social media platforms have become an important conduit and rich source of data for assessing public attitudes and behaviors during health emergencies. In light of the lockdowns and restrictions imposed due to the COVID-19 pandemic, social media platforms have emerged as key forums for the public to express their opinions and experiences pertaining to the pandemic and vaccinations. Examination of social media data could reveal significant trends, patterns, and changes, and can thus serve as a tool for health surveillance and monitoring the trends. This study builds upon the extant infoveillance research on the COVID-19 pandemic by focusing on the discourse pertaining to COVID-19 vaccinations in Twitter. We analyzed over 2.94 million tweets from January 1, 2021, to April 30, 2021, to explore the trends, sentiments, and key themes pertaining to COVID-19 vaccinations.
There is growing interest in understanding public attitudes and opinions about COVID-19 vaccinations. Studies have found vaccine hesitancy to be prevalent globally across multiple countries, although there is some preliminary evidence about lower levels of hesitancy in lower- and middle-income countries as compared to developed nations such as the United States [
The extant studies have collectively helped us to uncover some key public concerns and trends regarding vaccinations, vaccine advocacy, and hesitancy. However, most of the existing studies have used data from early periods of the COVID-19 pandemic or initial phases of vaccination. Some of these studies have also not differentiated if the source of a tweet is an individual or an organization. Several thousands of tweets are typically made by news outlets, health agencies, or other organizations. From an infoveillance perspective, it is critical to examine the social media discourses pertaining to COVID-19 vaccines by the common public rather than by news agencies or other organizations. Building upon the emerging body of research, our study differs from this prior research in the following ways. First, we focused on tweets made between January and April 2021, capturing public attitudes during active periods of vaccinations in many countries. Second, we examined English-language tweets from all over the world, without restriction to a region or a country. Third, we focused on tweets made by individuals only, thus capturing public sentiments and concerns. Our study is uniquely positioned and differs from many other similar studies listed in
Summary of key studies on COVID-19 vaccines using social media data.
Source | Data set | Time period | Key findings | Limitations/remarks |
Yin et al [ |
1.75 million Weibo messages from China | January to October 2020 | Identified public opinions pertaining to pricing, side effects, and inactivated vaccines | Restricted to Chinese-speaking Weibo users, including residents of China and those living abroad. The study used posts from verified users. |
Hussain et al [ |
23,571 Facebook posts from the United Kingdom and 144,864 from the United States; 40,268 tweets from the United Kingdom and 98,385 from the United States | March 1 to November 22, 2020 | Overall averaged positive, negative, and neutral sentiments were at 58%, 22%, and 17% in the United Kingdom, in contrast to 56%, 24%, and 18% in the United States, respectively. Public optimism regarding vaccine development, effectiveness, clinical trials, concerns over their safety, economic viability, and corporation control were identified. | Geographical scope included the United Kingdom and the United States. The study does not mention excluding tweets made by organizations and news outlets. |
Guntuku et al [ |
4 million tweets originating from 2957 US counties | December 1, 2020, to February 28, 2021 | Topics identified include side effects, conspiracy theories, trust issues in the US health care system in December 2020; mask wearing, herd immunity, natural infection, and concerns about nursing home residents and workers in January 2021; and access to black communities, vaccine appointments, family safety, and online misinformation campaigns in February 2021. Geographic variations on the topics across different counties were also identified. | Geographical scope was restricted to the United States. The study does not mention excluding tweets made by organizations and news outlets. |
Bonnevie et al [ |
1,438,251 tweets; 6498 per day | Antivaccine tweets from February 15, 2020, to June 14, 2020, as compared to those in the pre-COVID-19 period of October 15, 2019, to February 14, 2020 | Mentions of vaccine opposition increased by 79.9%. The themes identified were negative health impacts, pharmaceutical industry, policies and politics, vaccine ingredients, federal health authorities, research and clinical trials, religion, vaccine safety, disease prevalence, school, and family | No mention of exclusion of tweets made by organizations and news outlets |
Griffith et al [ |
3915 tweets about vaccine hesitancy from Canada | December 10, 2020, to December 23, 2020 | Vaccine hesitancy was attributed to the following themes: concerns over safety, suspicion about political or economic forces driving the COVID-19 pandemic or vaccine development, a lack of knowledge about the vaccine, antivaccine or confusing messages from authority figures, and a lack of legal liability from vaccine companies | Geographical scope restricted to Canada, with limited sample size; manual coding of tweets |
Hou et al [ |
7032 tweets and Weibo posts from five locations: New York, London, Mumbai, Sao Paulo, and Beijing | June and July 2020 | Beijing users (76.8%) had a higher vaccine acceptance rate as compared to those in New York (36.4%). Concerns expressed included: vaccine safety, distrust in governments and experts, widespread misinformation, vaccine production and supply, vaccine distribution, and inequity | Manual coding of tweets and Weibo posts from five locations, with limited sample size. However, this study excluded posts from news outlets and organizational accounts |
Yousefinaghani et al [ |
4,552,652 tweets about COVID-19 vaccines | January 2020 to January 2021 | Sentiment analysis revealed positive being the dominant polarity and having higher engagement. Themes among the positive-sentiment tweets were happiness and hope, support, and religion. Themes among the negative-sentiment tweets were fear and frustration, disappointment, anger, and politics. More discussion on vaccine rejection and hesitancy as compared to provaccine themes | Examined tweets from six countries: the United States, the United Kingdom India, Australia, Canada, and Ireland. No mention of excluding organizational tweets. |
Hu et al [ |
308,755 geo-coded tweets from the United States | March 1, 2020, to February 28, 2021 | Identified three phases along the pandemic timeline and documented changes in public sentiments and emotions. An increase in positive sentiment coupled with a decrease in negative sentiment concerning vaccines were noted in most states. Major international or social events and announcements by influential leaders or authorities associated with changes in public opinions toward vaccines. | Geographical scope restricted to the United States. No mention of excluding organizational tweets |
Lyu et al [ |
1,499,421 tweets | March 11, 2020, to January 31, 2021 | 16 topics under five broad themes were identified: opinions and emotions around vaccines and vaccination, knowledge around vaccines and vaccination, vaccines as a global issue, vaccine administration, and progress on vaccine development and authorization | Did not exclude organizational tweets, but eliminated tweets by bots and fake accounts |
Eibensteiner et al [ |
Poll of 3439 Twitter users | February 12, 2021, and February 19, 2021 | 45.9% of Twitter users felt the safety of the COVID-19 vaccines to be adequate; over 82.8% responded affirmatively about taking the vaccination | Used an anonymized polling/survey method with a limited sample of Twitter users |
In this research, we sought to uncover important themes underlying the social media discourse pertaining to COVID-19 vaccinations. This will help us to better understand how individuals feel about COVID-19 vaccinations, their inclinations for uptake, as well as reasons behind their hesitancy. Given the prevalence of vaccine hesitancy worldwide [
Our specific research goals were to (1) explore the themes and topics underlying social media discourse pertaining to COVID-19 vaccines and (2) uncover trends and temporal variations in sentiments underlying COVID-19 vaccine discourse in Twitter.
This study used publicly available and accessible tweets made by individuals on the Twitter platform, which formed the data set used for our analysis. We present our analysis in aggregate form without identifying specific individuals who made the Twitter posts. Therefore, the activities described do not meet the requirements of human subjects research and did not require review by an institutional review board.
We used the Python scraper
We used a machine learning approach to separate tweets made by individuals and organizations. Following the approach outlined by Chandrasekaran et al [
Our next step involved preprocessing and cleaning of tweets using a set of libraries in Python. Using the
Topic modeling is an unsupervised machine learning method for identifying latent patterns of words in a large collection of documents. The most representative method for topic modeling is latent Dirichlet allocation (LDA), which is a generative probabilistic method [
We used CorEx and iterated with a varying number of topics (eg, 5, 10,15, 20, 30). The keywords for different topics were assessed by the authors to ascertain their coherence and meaningfulness pertaining to a topic. The total correlation scores were compared across iterations to decide on the optimal number of topics produced. Next, we reviewed the results to infer appropriate topics on the basis of keywords. We also examined a set of randomly chosen tweets for each topic to assess if those tweets were consistent with the topic. Through discussions, the authors then grouped the topics into broader themes. Our procedures are consistent with similar studies that have examined social media data using text mining and topic modeling [
In addition to topic modeling and sentiment analysis, we also performed qualitative analysis of tweets in each theme/topic to obtain further insights and temporal trends in the vaccine-related tweets.
Our data gathering resulted in an initial set of 3,707,187 tweets. We removed 762,657 tweets made by organizations. Consistent with our research goal of assessing public sentiments and attitudes, 2,944,530 tweets made by 1,210,225 Twitter users were included in our analysis.
The trends in the number of tweets about COVID-19 vaccines from January to April 2021 are presented in
Proportion of COVID-19 vaccine–related tweets from January to April 2021.
Our CoreX topic modeling resulted in 16 topics (
Topics and broad themes underlying COVID-19 vaccine–related tweets (N=2,944,530).
Themes and topics | Tweets, n (%) | |
|
508,658 (17.27) | |
|
Vaccination disclosure | 201,102 (6.83) |
|
Postvaccination symptoms and effects | 307,556 (10.44) |
|
462,529 (15.71) | |
|
Vaccine efficacy | 139,280 (4.73) |
|
Clinical trials, approvals, and suspensions | 182,673 (6.20) |
|
Vaccine distribution and shortage | 140,576 (4.77) |
|
630,606 (21.42) | |
|
Vaccine affordability | 116,205 (3.95) |
|
Regulation: mandatory versus optional | 410,466 (13.94) |
|
Travel | 103,935 (3.53) |
|
176,329 (5.99) | |
|
Vaccination appointment and scheduling | 105,586 (3.59) |
|
Vaccination sites | 70,743 (2.40) |
|
1,093,050 (37.12) | |
|
Vaccination eligibility and policies | 76,605 (2.60) |
|
Vaccination promotion and advocacy | 264,368 (8.98) |
|
Vaccination hesitancy | 371,843 (12.63) |
|
Opinion leaders and endorsement | 172,002 (5.84) |
|
Hoax/conspiracy | 208,232 (7.07) |
Gratitude toward health care workers | 73,358 (2.49) |
We computed the sentiment scores of COVID-19 vaccination tweets and tracked their changes over the time period of our study. The results are presented in
We further examined the trends in sentiments of the 16 topics over time. These results are presented in
We also examined the trends in the average sentiment score for each of the 16 topics over the time period of examination and plotted the average compound scores by topic and week. The results are presented in
Proportions of positive, negative, and neutral tweets about COVID-19 vaccination.
To further examine the public sentiments and attitudes toward COVID-19 vaccines and vaccination rollouts, we qualitatively examined the tweets for the top three themes that emerged from our topic modeling assessment.
Approximately 14% of the tweets about COVID-19 vaccination in the study period focused on the issue of whether vaccines need to be made mandatory. Many tweeters argued for mandatory vaccination, especially in places of work, schools, education institutions, and for travel:
Just like having a vaccination card to go to school, I feel businesses and all schools should make it mandatory to have Covid vaccine
Would you refuse to take the Covid vaccine; if it became compulsory to work?
If, eventually, we need to show proof of vaccination to go to theatres, restaurants, sporting events etc then no, it’s not truly optional - by any reasonable measure that’s coerced vaccination.
Tweeters also argued for making COVID-19 vaccines mandatory to health care workers. Several countries such as France have introduced mandatory vaccination requirements for health care workers. Saudi Arabia announced that all of the employees in the public, private, and nonprofit sectors must be vaccinated before they can return to work. Italy introduced a vaccination requirement for all of their health care workers and pharmacists [
I support #MandatoryVaccination for nurses
Let’s keep pushing for #MandatoryVaccination of those who care for our most vulnerable Ridiculous that we're making vaccination optional for healthcare workers...vaccinate or GTFO.
Tweeters opposed to mandatory vaccination opined about how such mandates can be extended to other areas and expressed displeasure:
Its all part of the #mandatoryvaccination by coercion agenda. They are going to achieve it by: Divide and Rule -> getting the #vaccinated to blame the #unvaccinated. Threatening people with no sport events pubs etc. These narratives will grow and grow over the coming months. What happens to #MyBodyMyChoice if we’re forced into #mandatoryvaccination ? Next it will be #forced #abortion and #sterilization?
Approximately 12.63% of the tweets in our data set were about vaccine hesitancy that highlighted the reluctance of a set of Twitter users to receive COVID-19 vaccines. When we qualitatively examined these tweets, we found tweeters simply spelling out their stance to reject the vaccines, with many users highlighting reasons for not accepting vaccines. Promoting COVID-19 vaccines will need a clear understanding (particularly for those against COVID-19 vaccines) of whether people are willing to be vaccinated and the reasons why they are willing or unwilling to do so. We observed some common reasons cited by Twitter users for their vaccine hesitancy. Some users expressed concerns on how quickly the vaccines were developed and wondered about safety. For instance, one user tweeted “I don't trust a vaccine that was developed in such a short period of time, when we can’t even find one for so many other illnesses,” and another user tweeted “I don’t trust that jab...it’s usually years before a vaccine is ready....too rushed.. I don’t trust it.” There were others who expressed concerns about effectiveness of vaccines and if the vaccines can protect against newer strains of the virus. As one tweeter stated, “I’m not getting the vaccine. No one knows what’s in it or the long term effects of it, or if it can stop new variants.” From some other tweets, we observed public mistrust of the pharmaceutical industry, medical community, and governments:
I don't trust pharma and I won't be having any covid vaccine till it's been around for a while longer and the guinea-pigs have put it to good testing
I don’t trust this vaccine, I don’t trust the CDC, I don’t trust free donuts from Krispy Kreme (LMFAO), i don’t trust our government
Nope! Not getting the “vaccine”. I don’t trust the government nor companies who work with the government
Over 10% of tweets in our data set were about users sharing their experiences on symptoms and side effects of COVID-19 vaccines. Moreover, the average compound sentiment for this topic remained negative throughout the 4-month period. Twitter users shared information about the dose and their experiences subsequent to vaccination. While some users reported little or no side effects (“24 hours after my first jab of the Covid-19 vaccine, I have not observed any untoward effect from the vaccine”), others provided more detailed information on side effects and how they progressed over a period of time following the vaccination:
Had the jab at 11am yesterday and the chills & aches started at about 7pm last evening. Lots of Tylenol & fluids.
I received my 2nd covid shot yesterday morning. The biggest side effects were weakness and terrible dizziness.
Day 2 post-vaccine was no cake walk. Fever, major aches, brain fog, sore everywhere. But man am I glad I got it
Mentions of side effects were often accompanied by messages expressing elevated feelings about protection against the virus:
I had side effects from the vaccine, but that 24 hours of chills and fever was worth it to keep myself, friends, family, and my community safe.
I would much rather take 48 hours of aches and chills from the second dose of the vaccine than risk gasping for my last breath in an ICU away from family.
A growing number of studies have used data from social media to explore and understand public concerns and attitudes about the COVID-19 pandemic. As governments around the world are trying to tackle the pandemic through mass vaccination, it is important to uncover public opinions and attitudes toward COVID-19 vaccines. We used a repository of approximately 3 million tweets from January 2021 until the last week of April 2021 to uncover the trends in sentiments of various themes and topics pertaining to COVID-19 vaccines. We focused on tweets made by individual users and excluded those made by news outlets and other organizations. Through topic modeling, we found 16 topics pertaining to COVID-19 vaccines that were grouped into six broad themes. Further, we examined sentiments associated with these topics and the changes in sentiments over the 4-month period.
A key finding from our study is that the regulation pertaining to COVID-19 vaccines was the most discussed issue by Twitter users. The number and proportion of tweets on this theme were greater than those for all the other topics. The proportion of tweets with positive sentiments about regulation of the vaccination outweighed the proportion of negative and neutral tweets pertaining to this topic. We found vaccine hesitancy to be the second most discussed topic. We also observed negative sentiment scores for many weeks for this topic. Our qualitative analysis provided some preliminary insights into reasons behind vaccine hesitancy: shorter duration of the vaccine development cycle, concerns about effectiveness of the vaccine in controlling the virus and its variants, and general mistrust about the pharmaceutical and medical industries and governments. Another topic that was widely discussed was postvaccination side effects and symptoms. The average sentiment scores for this topic were negative throughout the time period examined.
To control the COVID-19 pandemic, it is important that a substantial portion of the worldwide population acquire immunity through vaccination. Policymakers and public health officials are increasingly focusing on ways to boost and accelerate vaccine uptake. Vaccination campaigns are being designed to address misinformation and public concerns regarding the vaccines. In addition, several efforts are being made to increase vaccine supply, introduce incentive mechanisms for encouraging vaccine uptake, and enhance public education and outreach programs. However, our findings indicate that vaccine mandates and vaccine hesitancy continue to dominate the minds of the general public, as can be seen from their posts on social media. It is important to take their attitudes into account while framing and designing vaccination campaigns and programs.
It should also be noted that most COVID-19 vaccines have been approved for emergency use and authorization, rather than through a regular licensing route. As more vaccines that are currently authorized for emergency use obtain regular approval and licenses by authorities such as the FDA, the issue of vaccine mandates is likely to gain more prominence. More employers and authorities could enforce vaccine mandates. Schools and educational institutions in many parts of the world have started mandating COVID-19 vaccines. Further, vaccination is also a requirement for most international travel. It is more likely to become a requirement for even domestic travel in several countries. A complementary approach to mandating COVID-19 vaccines is creation of trust and favorable attitudes toward vaccines in the minds of the public. Mass outreach and education programs along with incentives for vaccination can go a long way in accelerating vaccination uptake. Further, endorsement by leaders and celebrities and experience-sharing by peer individuals could also help alleviate concerns regarding vaccines.
This study points to the key issues surrounding COVID-19 vaccinations in the minds of the general public, as expressed through social media. Findings from our study bear important implications for the design of vaccination campaigns and programs. Identification of reasons for vaccine hesitancy throws light on questions that need to be answered by health policymakers and health care practitioners in order to allay the apprehensions pertaining to vaccines and their side effects. Moreover, experience sharing from the public on vaccination, side effects, and their mindsets could also serve as a morale booster for others. Some social media posts also serve as testimonials for the efficacy of vaccinations and their effectiveness. Future vaccination drives and campaigns can take into account the experiences of a fairly large body of individuals to design appropriate responses to increase vaccination uptake.
This study used tweets posted from January 1 to April 31, 2021. Vaccination efforts accelerated in several parts of the world shortly after (June-July of 2021), which have not been captured by our study. It should also be noted that we used a machine learning classifier to separate tweets made by individuals and exclude those made by organizations and news outlets. This helped us to remove numerous tweets made by media outlets and organizations so that we could capture the attitudes of the general public. The classifier exhibited an accuracy of 91.81%, which is comparable or better than those reported in many other studies [
Another limitation is that we covered only tweets posted in the English language. Due to the nature of the data we gathered, we did not explore any geographical disparities in the tweets, which could also be a fruitful extension to our work. Another extension of our work would be to examine emotions expressed in tweets pertaining to COVID-19 vaccinations. Another important limitation of our study is that we have captured only the attitudes and opinions of Twitter users, who have a presence in social media. Twitter users tend to be technology-savvy, adept in using social media, and own smartphones, and therefore may not represent the entire population set. A larger set of the population who do not have a presence on Twitter has not been covered by our study.
With variants of the virus causing COVID-19 creating multiple waves of the pandemic in several countries, it is important to accelerate the rate of vaccinations and improve uptake. As COVD-19 vaccination efforts move forward, it will be important to continue to monitor public opinions regarding vaccine mandates, vaccine hesitancy, and vaccination uptake. Some individuals and groups are likely to continue to oppose vaccines, whereas there may be many others who could be convinced by appropriate education and outreach programs. While mandates by governments or employers could be contested on legal grounds, appropriate exemptions will need to be made for people with certain health conditions or special situations. Infoveillance based on social media data can provide rich insights for policymakers and health officials to frame appropriate policies and programs for COVID-19 vaccination.
Themes, topics, and keywords.
Trends in sentiments of tweets.
Average sentiment scores and trends.
Correlation Explanation
Food and Drug Administration
latent Dirichlet allocation
valence aware dictionary and sentiment reasoner
None declared.