🔍 Real-Time Sentiment Analyzer

Instantly classify text as Positive or Negative using DistilBERT — no API key, no cloud cost, runs on any machine.

---

Features

• Single-text analysis — paste any text and get an instant POSITIVE / NEGATIVE label with a confidence score and visual bar
• Batch CSV processing — upload a CSV with a text column and analyze hundreds of rows at once
• Downloadable results — export your batch results as a clean CSV file
• Sentiment distribution chart — auto-generated bar chart showing the split across your dataset
• Zero cost — powered by a free HuggingFace model that runs fully on CPU; no GPU or paid API required
• Clean Gradio UI — two-tab layout built with gr.Blocks and gr.themes.Soft()

---

Quick Start

git clone https://github.com/your-username/sentiment-analyzer.git && cd sentiment-analyzer
pip install -r requirements.txt
python app.py

Then open your browser at http://localhost:7860

---

How It Works

DistilBERT is a smaller, faster version of Google's BERT language model — trained by reading billions of sentences from the internet until it learned what words, phrases, and sentence structures tend to mean. It was then fine-tuned specifically on movie and product reviews so it knows the difference between language that expresses satisfaction and language that expresses frustration. When you give it a sentence, it converts every word into a stream of numbers, passes those numbers through 6 layers of "attention" math that weighs how each word relates to every other word, and produces a single prediction: Positive or Negative, along with a confidence score showing how certain it is.

---

Demo

— add your screen recording here

---

Project Structure

sentiment-analyzer/
├── app.py                # Gradio Blocks UI + tab logic
├── model.py              # Lazy model loading + analyze_text / analyze_batch
├── requirements.txt      # Pinned dependencies
├── sample_reviews.csv    # 10 Amazon-style test reviews
└── README.md

---

Use Cases

• E-commerce — automatically flag negative product reviews for urgent review
• Social media monitoring — gauge public sentiment on brand mentions or campaigns
• Customer support — prioritize tickets that contain strongly negative language
• Research — quickly label large text datasets without manual annotation
• Education — learn how transformer-based NLP pipelines work end-to-end

---

Deploying to HuggingFace Spaces

1. Create a new Space at huggingface.co/spaces (choose Gradio SDK)
2. Push this repository to the Space's git remote
3. HuggingFace will auto-install requirements.txt and run app.py

Or for a quick public share link, change the last line of app.py to:

demo.launch(share=True)

---

License

MIT © 2024 — free to use, modify, and distribute.