The Coronavirus Tweets Text Mining and Sentiment Prediction

Background

Coronavirus disease 2019 (COVID-19) is a contagious disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The first case was identified in Wuhan, China, in December 2019[6]. The relevance of the COVID-19 global pandemic in Mar, 2020 has raised the attention of researchers all over the world.

Since there wasn’t any proved effective medicine or vaccine back then, CDC initiated a quarantine guideline for COVID-19 prevention. Most of the schooling and work were moved online until now. Social Media became an essential resources for people to communicate with each other. It certainly made our lives convenient. However, it’s a double-edged sword. Misinformation, fake news and extreme emotion can mislead the public. In a way, remote work and learning increased the anxiety in public because of the social media.

One of largest online public communities Twitter is an important resource for people to learn about this new virus and communicate with each other. It’s a very significant to detect the negative sentiment, rumors and misinformation from those tweets to avoid the public anxiety and the topic trending during the pandemic. This research is meant to be a social behavior study focusing on the Coronavirus Tweets Text Mining and Sentiment Analysis and will be flexible to be transferred to other social media text datasets

representing and communicating the information with the public.

Flask App

check out the Flask app in the Flask app folder. A quick peak of the app

Fig.0 A demo

Fig.1 Word Cloud and bar charts line charts

Fig.2 The Bert Fine-tuned models usage for sentiment

Also deployed version here

see the details at https://bis634.herokuapp.com/