Public Perception of COVID-19 Vaccine by Tweet Sentiment Analysis

Abstract

The COVID-19 pandemic has given a tremendous impact on the economy, changed the way of living of the whole world. Many lives are lost, the labor markets are affected and the people lifestyle are changed in order to limit the impact of COVID-19 pandemic. As of now, it seems that the COVID vaccine is the only solution for the world to be safe again. In the United Kingdom and the United States, many studies have been conducted on the sentiment analysis where the emotions of participants before vaccination and after vaccination are observed. The first batch of vaccines has been launched at the end of 2020 while some developed countries started early vaccination campaigns, and others are still in the process of ordering vaccines and remained unvaccinated until early 2021. The vaccinations are prioritized on the high-risk groups, such as medical workers and the elderly population. Vaccination for people under age 18 are still not available in the initial stage. Despite the executions of vaccinations, there are various opinions on whether the COVID-19 vaccines are safe, and a number of the population remain skeptical of taking the vaccine. In this research, we analyze tweets to understand public perception on the COVID-19 vaccine by classifying the sentiments and attitudes towards the vaccination and the available types of vaccine [1]. Social media is an appropriate source of research to analyze public attitudes towards COVID-19 vaccine and what they feel about the various brands of the vaccine in the market. For this research, tweets written in two languages, English and Japanese, have been collected. In Japan, some related surveys on public willingness for vaccinations and the sentiment analysis are already conducted. This study randomly surveys on the users’ tweets about COVID-19 vaccination and their emotions expressed in their tweets. Due to the certain vaccination accidents, people in various countries become more concerned about the side effects and safety of the vaccine due to local deaths of various circumstances and unknown causes. In an attempt to help assess and understand public sentiment towards the initial stage of the vaccination campaign, sentiment analysis tools are utilized. It can discover that there are different sentiment patterns observed in different regions and time points as well as in different vaccine brands. It is expected that the text categorization process will be conducted using NLTK’s dedicated Twitter corpus. In social media data, users enter multiple punctuation marks, acronyms and emotions to express their emotions. TextBlob tool will be used for analysis, which computationally identifies and classifies text into three emotions: positive, negative or neutral. TextBlob is used because it processes data by including various letters, symbols, etc. in its dictionary. In this method, each word in the dictionary is based on whether it is positive or negative, while adding an emotional analysis of commonly used expressions. In this way, people’s attitudes towards vaccines in the UK, the US and Japan can be analyzed. The accuracy of the method is 73.3% in English and 71.9% in Japanese. The results show that the British and Americans are more neutral and positive about vaccines, while the Japanese are more pessimistic about vaccines [2].