Market Sentiment Dashboard

A real-time dashboard for analyzing sentiment in financial news articles related to publicly traded companies. Built with FastAPI (backend NLP API) and Dash (frontend). Deployed via Docker Compose.

---

🔍 Features

• NLP-based sentiment analysis (positive, neutral, negative)
• Highlights most confident and most polarized news articles
• Real-time dashboard interface (Dash)
• REST API with FastAPI

---

⚙️ Stack

• Python (FastAPI, Dash, transformers)
• Docker Compose (2 services: api and dashboard)
• News ingestion via API (Finnhub)

---

🚀 Quickstart (Development)

1. Clone the repository

git clone https://github.com/LouisSch/Market-Sentiment.git

2. Create a .env file

FINNHUB_API_KEY=your_api_key_here

3. Launch the services

docker-compose up --build

• The API will be available at: http://localhost:5002
• The dashboard will be available at: http://localhost:8050

⚠️ This setup uses local volumes for live editing (development mode). For production, use COPY in Dockerfiles and avoid volume mounting.

---

🖼️ Preview

---

🧠 Motivation

This project was built to visualize and explore the dominant sentiment around equities based on media coverage. While not predictive, the tool offers insights into the perception dynamics that may influence markets.

---

🔗 Links

• 🌐 Personal Website: https://lschirra.dev

---

📎 License

MIT License