The widespread COVID-19 pandemic imposed enormous burdens and severely upset societies and economies worldwide. Medical institutions and healthcare professionals diverted all their efforts towards developing vaccines to combat the virus.
While many medical institutions and pharmaceutical industries were successful in developing a vaccine, people were still cynical about the effectiveness of those vaccines. The speed at which these institutions developed the vaccine made people question the quality of the vaccine. The new emerging side-effects of each only fueled the concerns about the vaccines.
In this case study, we will examine all tweets related to covid vaccines using sentiment analysis to evaluate the public sentiments towards the COVID-19 vaccines.
In March 2020, the WHO declared the COVID-19 virus a global health crisis. We are 2021 now and the conditions are better, but some parts of the world are still on lockdown and new strains of the virus have emerged. Vaccinating the general public is one of the best ways to halt the spread of the virus. However, growing human rights concerns, anti-vaccine movements, and doubt about the effectiveness of the vaccine have slowed down the mass vaccination drives in many countries. The citizens would rather wait and make sure that there are no side effects of the vaccine before getting vaccinated.
The COVID-19 vaccination is currently a highly debated topic on social media platforms and news. In this research, we use Twitter data to analyze and evaluate the public sentiments towards the COVID-19 vaccination. The number of confirmed and presumptive positive cases has reached over 30 million in the US. There are also over 539 thousand deaths reported from the overall cases.
While the risk still remains high, the number of daily cases is decreasing in 2021. However, underlying health conditions and new strains of the Coronavirus can change the situation at any time. Further, the recent news about the blood clots forming in some people after being vaccinated with the AstraZeneca vaccine forced many countries to halt their vaccination drives.
To conduct successful vaccination drives, Federal and State governments have to first understand the major concerns of the general public. Only then can they implement effective measures towards eradicating bias and misinformation regarding COVID vaccination. This research will help government officials and policymakers understand the public sentiment and assist in planning effective measures to conduct successful mass vaccination drives.
The outbreak of the COVID-19 pandemic and the vaccination attempts of various governments across the globe has been a fairly discussed topic. But what’s the most discussed topic among the general public is the effectiveness of the vaccines and their side effects.
Social media platforms are the most used channels by the general public to discuss issues related to COVID and vaccination. The purpose of this research is to understand the sentiments of the general public in the US. As Twitter attracts the most complex and diverse conversations about the topic and is widely used by American citizens, we chose to conduct our research using Twitter data.
The main objective of this study is to understand the attitudes and sentiments of the general public in the US towards COVID vaccination. We would also focus on identifying the most popular topics of discussion. We will be performing this research using BytesView’s various text analysis techniques.
Research data extraction and pre-processing:
Data Sources: Twitter data scraped using the Twitter API
Data Type: COVID Vaccines related tweets
Keywords used to extract data:
- Pfizer Vaccine
- Moderna vaccine
- COVID vaccine
Twitter data scraping time frame
The analyzed tweets were scraped between 3rd March and 6th April.
The volume of the research data
1 Million COIVID vaccine-related tweets
Research Data Pre-Processing
The research data pre-processing involves:
- Text Clean Up: This part of the process involves removing all the unnecessary information from the research data such as ads or unwanted characters. It also involves standardizing data from binary information sources.
- Tokenization: This step of the process involves splitting the text from the research data into white spaces. It breaks down textual data into words and sentences, also referred to as tokens.
- Parts of Speech Tagging: Just as the name suggests, this step of research data pre-processing involves tagging every word from the textual data with the right parts of speech.
Text analysis models applied:
Following are the text analysis models used to extract data or insights from complex and unstructured Twitter data.
1. Sentiment Analysis: Sentiment analysis is a machine learning and NLP-based text analysis technique that helps machines interpret human language and analyze the sentiments expressed in the text. BytesView’s sentiment analysis solution analyzes and classifies text data into three types:
We will be using the sentiment analysis solution to examine and classify tweets based on the positive, negative, or neutral sentiment expressed in the text. This will help evaluate the overall public sentiment of the US citizens towards the Pfizer and Moderna vaccines.
2. Topic labeling: Topic labeling is a text analysis model that is frequently used to analyze unstructured textual data to identify popular topics of discussion. This model is capable of labeling unstructured feedback data. Each label has a pre-defined set of keywords. The analysis model recognizes the presence of these keywords in textual data and analyses the text’s theme. These pre-defined labels are then used to categorize the unstructured textual data.
We will be using the model to identify the most discussed topics related to the Pfizer and Moderna vaccines on Twitter.
3. Keyword extraction: Keyword extraction is a text analysis technique that automatically extracts the most used keywords and expressions from large volumes of unstructured textual data.
We will use the keyword extraction model to automate the identification and extraction of the most used keywords in the tweets related to the Pfizer and Moderna vaccine posted the US citizens.
4. Emotion Analysis: Emotion Analysis is a text analysis technique that enables you to analyze and classify unstructured textual data based on the emotions expressed by the author. BytesView emotion analysis solution can classify textual data into the following types:
We will use emotion analysis to classify tweets as per the emotion expressed in the text. This will help in better understanding the thoughts of the general public when it comes to Pfizer or Moderna COVID-19 vaccines.
The result of the analysis will be discussed in two phases. The first phase solely focuses on the analysis of tweets mentioning the keywords Pfizer and Moderna. The second phase focuses on the analysis of all 1 million tweets related to COVID-19 vaccination.
Phase 1: Analysis of tweets related to Pfizer and Moderna Vaccines
The first phase of the analysis includes the following:
Pfizer VS Moderna Sentiment Analysis
From the analysis, we can conclude that the Pfizer vaccine 63% of the Twitter users sent out tweets with negative sentiments while 15% of the users were tweeting positive sentiments. Lastly, the remaining 21% of users expressed neither positive nor neutral sentiments and stayed neutral in their opinion.
As for the Moderna vaccine, 57% of the tweets lie in the negative category of sentiments. 23% of the tweets express positive sentiments for the vaccine and 19% of the tweets expressed neutral sentiments.
The analysis states that most of the users tweeted negative sentiments for both the vaccines. Although, tweets related to Pfizer were more negative when compared to the tweets for the Moderna vaccine.
Pfizer VS Moderna Emotion Analysis
Now that we have analyzed the sentiments related to both the vaccines, let’s dive a little deeper and analyze the emotions expressed by the users.
After analyzing the tweets, we can conclude that most tweets expressed happy (joyful) emotions (41%) for Pfizer. It was followed by fearful emotions (31.48%) as people were still concerned about the effectiveness and side effects of the vaccine. The other sentiments show similar results and can be viewed in the figure above.
As for tweets related to Moderna, most tweets expressed fearful emotions (46.70%) as news related to the side effects and effectiveness of the vaccine on mutated strains were a popular topic at the time. The second dominant emotion expressed in the tweets is happiness. It comprises over 30% of the total tweets. You can check the statistics for the other emotions in the above figure.
Phase 2: Analysis of tweets related to COVID vaccination
The second phase of the analysis includes the following:
COVID vaccination tweets sentiment analysis
The most dominant sentiment expressed in the tweets is negative. Over 59% of the tweets express negative feelings about vaccination. This may be due to the identification of the side effects of various vaccines, new virus strains, and doubts about their effectiveness. As for the rest, both positive and neutral sentiments display very similar statistics.
COVID vaccination tweets emotion analysis
For emotion analysis, the most prevalent emotion expressed in the tweets is fear. Over 53% of users expressed fear in their tweets followed by happiness (23%), sadness (7.5%), anger (9%), and neutral emotion (7%). The least expressed emotion in the analyzed tweets was love (0.2%).
Most used keywords
The above word cloud includes references to the most used keywords in the tweets. Through this, we can identify the most discussed topics related to COVID vaccination.
In this particular dataset, as expected, the most discussed topics are the vaccines themselves (effectiveness, side-effects, availability, approval, etc). People also discussed vaccine research-related topics such as human trials and first dose. Apart from COVID people also discussed how COVID vaccine mRNA technology can be used to create effective HIV vaccines.
The study was conducted to analyze the overall sentiments of US citizens in relation to COVID vaccination. The study shows people expressing more positive sentiments for the Pfizer vaccine in comparison to the Moderna Vaccine. As for emotions, fear and happiness were the most prevalent kind. The most expressed emotion for the Moderna vaccine was fear. As for the Pfizer vaccine, the most expressed emotion was happiness.