This project aims to classify vaccine sentiments using BERT and GloVe embeddings. The dataset used for this project is the "Global COVID-19 Twitter dataset," which contains tweets from Australia, India, Brazil, Indonesia, Japan, USA and UK.
The project utilizes the dataset from Kaggle:
Janhavi Lande, Yashwant Kaurav, Cathy Yu, & Rohitash Chandra. (2022). Global COVID-19 Twitter dataset. Kaggle. https://doi.org/10.34740/KAGGLE/DS/2397387
The repository includes individual Jupyter Notebook files for each country's dataset:
- Australia.ipynb
- India.ipynb
- Indonesia.ipynb
- Brazil.ipynb
- Japan.ipynb
These notebooks contain the filtered and preprocessed tweets for each country, and sentiment analysis has been performed on the tweets specific to that country.
Additionally, there is a combined analysis in the file EDA_Global_Covid_Dataset.ipynb. This notebook presents data analysis and visualization, combining the predicted sentiments from all countries' datasets.
To reproduce the results:
- Run the individual country's notebook to generate the predicted sentiment data.
- Input the generated sentiment data into the EDA_Global_Covid_Dataset.ipynb notebook to perform the combined analysis.
