News Analyzer — Summary, Sentiment & Political Bias Detector

A minimal web-based tool that fetches any news article URL, summarizes its core content, and analyzes both the sentiment and political bias using state-of-the-art NLP models.

Features

• Automatic article scraping via URL
• Text summarization using facebook/bart-large-cnn
• Sentiment analysis using distilbert-base-uncased-finetuned-sst-2-english
• Political bias classification with premsa/political-bias-prediction-allsides-DeBERTa
• One-click shutdown button (no need to hit Ctrl+C)

How It Works

Paste a URL from any valid news site.

Click Analyze.

Behind the scenes:

• Article is downloaded and cleaned
• Summarizer extracts key points
• Sentiment and political bias are detected from the raw article

Results are displayed in real time on the page.

Getting Started

1. Clone the repo

git clone https://github.com/shayan-sethi/newsanalyzer
cd newsanalyzer

2. Install dependencies

pip install -r requirements.txt

If newspaper3k gives issues, try:

pip install newspaper3k --force-reinstall

3. Run the app

python start_app.py

Then open http://127.0.0.1:5000 in your browser.

Models Used

• facebook/bart-large-cnn — summarization
• distilbert-base-uncased-finetuned-sst-2-english — sentiment
• premsa/political-bias-prediction-allsides-DeBERTa — bias classification