Sentiment Analysis and Topic Modeling from Tweets about the Covid-19 Vaccine

Abstract

Coronavirus disease 19 (COVID-19) was discovered near the end of 2019 in Wuhan, China, and has since spread rapidly throughout the world. The most effective technique for combating the COVID-19 pandemic is to rapidly and successfully develop COVID-19 vaccines. Most people's primary source of information on health and vaccination is now the Internet. Twitter is one such platform that allowed users to gather as well as disseminate information. This paper uses data mining techniques to extract data from Twitter and evaluates the sentiments associated with COVID-19 vaccines and vaccination drives being conducted in India. The purpose of the research was to gain information about the perceptions of the general public towards COVID-19 vaccines in India using Exploratory Data Analysis, Sentiment Analysis and Topic Modelling. Moreover, the research aimed to analyse the underlying factors that contributed to the respective attitudes among people. VADER Sentiment Analyzer was used to identify Positive, Negative and Neutral sentiments among people and the dominant sentiment came out to be Positive. Then, Exploratory Data Analysis was used to study different types of users with regard to the three sentiments and a time-based analysis was done to identify the events that triggered a particular reaction. To further analyse the public emotions, text2emotion was used and 5 different emotions (happy, sad, angry, fear and surprised) were identified. Using Latent Dirichlet Allocation various topics were identified related to the vaccines