Sentiment Analysis in Indian Sub-continent During COVID-19 Second Wave using Twitter Data

Abstract

In Indian sub-continent COVID-19 second wave started in early March 2021 and its effect was more lethal than the first wave, the confirmed cases and the death rate was higher than in the first wave. Unlike the national lockdown in 2020, this year different states have started imposing lockdown like restrictions spanning April-June 2021. This paper investigates the sentiments of the people using twitter messages during early period of the second wave. Two-weeks data is manually annotated and several machine learning models were built. The best performing models were used to predict sentiments for the next 2–3 weeks and analysis is presented. Predictions of public, commercial libraries were also analysed in the same context.