Demo Video: https://www.youtube.com/watch?v=lN1VXEbe8nU
Social media platforms like Twitter have become one of the most popular sources of information and communication, especially among young people. Millions of tweets are posted every day on various topics ranging from news, politics, sports, entertainment, and more. These tweets contain a wealth of information, including people's opinions, thoughts, and sentiments about various topics.
Therefore, our project aims to address these challenges by developing a system that can perform sentiment analysis and tag prediction on tweets accurately. We believe that this project can provide valuable insights to companies, individuals, and researchers by allowing them to make informed decisions and discover new information.
Our repository involes Git Large File (GLF), remember to install git-lfs in advance.
With streamlit installed, enter our frontend-streamlit directory, run:
streamlit run homepage.py
You can also check out our application on Streamlit Cloud with simply clicking the link below. streamlit cloud link
- data_preprocessing: Preprocess Twitter data for further use.
- frontend_streamlit: Our frontend generated by Streamlit with interactive dashboards.
- scraping: Consists of source code for data collection and scraped data.
- sentiment_analysis: Contains data cleaning for sentiment analysis, along with two approach for sentiment score.
- tag_prediction: Includes model related files for tag prediction, such as model training, input data cleaning, saved model, interface for model prediction, etc.
- Data Collection: Selenium
- Data Processing: Spark, Pandas, Numpy, NLTK
- Model Development: PyTorch, Scikit-learn, Transformers, Keras
- Data Visualization: Streamlit, Matplotlib, Plotly, Wordcloud
- Deployment: Streamlit Cloud
- Xiner Qian @qian-x2
- Xiaoxiao Duan @stelladuan
- Zeye Gu @Simonmon06
- Ziyao Cui @ziyaocui
- Jingyi Huang @huangj20, @xxxibdara