Financial News Sentiment Analyzer App is a user-friendly web application that analyzes financial headlines and returns sentiment probabilities (Positive / Negative / Neutral) using FinBERT (ProsusAI/finbert).
🌐 Live App: 👉 Click here to use the app
This project started as a prototype (Gradio) and was converted to a polished Streamlit web app for better interactivity and shareability. It provides quick sentiment insights for financial headlines — great for analysts, journalists, researchers, and investors.
Key features
- 🔍 Instant sentiment detection for financial headlines
- 🤖 FinBERT-based model trained/tuned for finance text
- 📊 Probability visualization for each sentiment class
- 💬 Clean, responsive UI built with Streamlit
- ☁️ Easy to run in Colab, CoreLab, Streamlit Cloud, or a server
This app uses ProsusAI/finbert from Hugging Face — a specialized BERT model for financial sentiment classification (Positive / Negative / Neutral).
- Framework: Streamlit
- Model: Hugging Face Transformers (FinBERT)
- Libraries: transformers, torch, scipy, pandas
- Runs on: Colab / CoreLab / Local / Streamlit Cloud