Detecting the Presence of COVID-19 Vaccination Hesitancy From South African Twitter Data Using Machine Learning

The COVID-19 vaccination campaign has faced significant challenges due to vaccine hesitancy, and social media platforms, especially Twitter, provide valuable insights into public sentiment. This project aims to detect COVID-19 vaccination hesitancy by analyzing South African Twitter data using machine learning algorithms. The system classifies tweets into three categories: negative sentiment (indicating vaccine hesitancy), neutral sentiment (representing factual updates or questions), and positive sentiment (supporting vaccination). Several machine learning models, including Decision Tree, Random Forest, and XGBoost, are employed for sentiment analysis. Each model is trained on a labeled dataset of tweets, and the system processes uploaded CSV files containing tweet data through a Flask-based web application. The application allows users to upload data and provides real-time sentiment classification with confidence scores. This system offers an effective approach for monitoring and addressing vaccine hesitancy by leveraging social media sentiment. The insights gained can aid health organizations and policymakers in creating targeted campaigns to mitigate vaccine hesitancy and improve public health efforts.

Keywords:

COVID-19, Vaccine Hesitancy, Sentiment Analysis, Machine Learning, Decision Tree, Random Forest, XGBoost, Social Media, Public Health, Public Opinion.