Twitter Sentiment Analysis
Duration: Aug 2023 – Oct 2023
Tools Used: Python, Pandas, NLTK, Matplotlib, TensorFlow, WordCloud, Tweepy
This project focuses on detecting signs of depression from tweets collected during the COVID-19 pandemic. We implemented Natural Language Processing (NLP) techniques to clean and analyze tweets, and applied deep learning models including CNN and LSTM to classify sentiment.
🔍 Highlights:
- Collected over 50,000 tweets using the Twitter API.
- Cleaned and preprocessed text using tokenization, stop word removal, and TF-IDF.
- Trained a CNN-LSTM hybrid model that achieved 87% precision and 79% recall.
- Visualized sentiment distribution using word clouds and pie charts.
- Created a confusion matrix and performance evaluation dashboard.
📊 Research Poster
🔗 GitHub Repository
Access the full code and documentation here:
View on GitHub⬅ Back to Projects