User Questions from Tweets on COVID-19: An Exploratory Study

TL;DR

This study uses NLP techniques to analyze COVID-19 related questions on Twitter, identifying key concerns and entities to inform policy and healthcare responses.

Contribution

It introduces an automated method combining topic modeling and NER to extract and categorize COVID-19 questions and entities from social media data.

Findings

Identified main COVID-19 questions on Twitter

Developed a NER model for key entities

Provided insights for policy and healthcare responses

Abstract

Social media platforms, such as Twitter, provide a suitable avenue for users (people or patients) concerned on health questions to discuss and share information with each other. In December 2019, a few coronavirus disease cases were first reported in China. Soon after, the World Health Organization (WHO) declared a state of emergency due to the rapid spread of the virus in other parts of the world. In this work, we used automated extraction of COVID-19 discussion from Twitter and a natural language processing (NLP) method based on topic modeling to discover the main questions related to COVID-19 from tweets. Moreover, we created a Named Entity Recognition (NER) model to identify the main entities of four different categories: disease, drug, person, and organization. Our findings can help policy makers and health care organizations to understand the issues of people on COVID-19 and it can be used to address them appropriately.