Preprocessed Twitter data using Natural Language Processing (NLP) techniques, including tokenization, part-of-speech tagging, and feature extraction. Developed and compared deep learning models (RNN, LSTM, and BERT) for sentiment analysis, with BERT achieving the highest accuracy of 89%.