DeepFeelings is project created to know the impact of a brand, or a product, on social media.
It differentiates between positive, neutral, and negative comments, in total and during a timeline, and it gives the most common topics surrounding it.
Work in progress
- Web scraping with Selenium to collect Amazon reviews
- Twitter API to collect tweets
- Model RoBERTa: for Sentiment Analysis
- Model Latent Dirichlet Allocation for clustering topics
- Data Visualization with matplotlib and plotly deployed on streamlit