This project is based on classifying the political leaning of twitter users based on their recent tweets. This is achieved by fine tuning the BERT-base model for classification using PyTorch. The end goal of this project is to cover all twitter users and analyze their exposure.
● Classified the political leaning of Twitter users based on their recent tweets with the accuracy of 80% which was achieved by fine-tuning the BERT-base model for classification using PyTorch
● Tweets are retrieved by a python script running on AWS instance and stored in MongoDB and prediction is stored using PostgreSQL in AWS instance