📊 Customer Comment Analyzer — Local AI with Ollama

Analyze customer comments about multiple products locally with AI — no internet required.

This app uses FastAPI + Gemma 3:1B (via Ollama) to read product reviews (Excel / CSV), summarize insights, and display each product’s summary beautifully on a modern web interface.

---

🚀 Features

• 📂 Upload Excel (.xlsx) or CSV files with product and comment columns
• 🤖 Analyze comments locally using gemma3:1b
• 🧠 Automatic grouping by product
• ⚡ Robust JSON handling (no crashes, even with malformed responses)
• 💎 Glassmorphic Persian UI
• 🖥️ Works fully offline

---

🧱 Tech Stack

Layer	Technology
Backend	FastAPI
Frontend	HTML + Jinja2 + Tailwind
AI Engine	Ollama (Gemma 3:1B)
Data	Pandas
Server	Uvicorn

---

⚙️ Installation & Run Guide

🪄 Step 1 – Clone the repo

git clone https://github.com/YOUR_USERNAME/CustomerCommentAnalyzer.git
cd CustomerCommentAnalyzer

🧩 Step 2 – Create a virtual environment

python -m venv .venv

Activate it:

Windows PowerShell

.venv\Scripts\Activate.ps1

Linux / macOS

source .venv/bin/activate

---

📦 Step 3 – Install dependencies

pip install -r requirements.txt

---

🤖 Step 4 – Install the AI model

Make sure Ollama is installed and running, then pull the model:

ollama pull gemma3:1b

---

🧠 Step 5 – Run the app

uvicorn main:app --reload

Visit → http://127.0.0.1:8000

---

📂 Project Structure

CustomerCommentAnalyzer/
├── main.py
├── templates/
│   └── index.html
├── static/
├── uploads/
│   └── .gitkeep
├── requirements.txt
├── .gitignore
└── README.md

---

📊 Example Dataset

product	comment
iPhone 13	عملکرد عالی و شارژ سریع
iPhone 14 Pro	طراحی زیبا ولی باتری ضعیف
iPhone 15 Pro Max	گران ولی فوق‌العاده قدرتمند

Save this as dataset.xlsx or dataset.csv and upload it in the app.

---

🧠 Local Model Notes

You can switch the model easily in main.py:

MODEL = "gemma3:1b"

Other working local models (if installed):

MODEL = "qwen2.5:1.5b"
MODEL = "deepseek-coder:6b"

---

🧾 License

MIT License © 2025 Hossein Fallah

---

🧭 Upcoming Updates

Here are 20 planned improvements already structured for GitHub milestones:

1. ✨ Add real-time progress bar during batch analysis
2. 📊 Add product sentiment score visualization (positive/negative)
3. 🧾 Export results as Excel or PDF
4. 💬 Multilingual support (English, Arabic, Persian)
5. 🧠 Choose model dynamically from UI
6. 🖥️ Dockerfile for easy deployment
7. 📦 Add REST API endpoint for programmatic access
8. 📈 Include token usage and model speed metrics
9. 🔐 Add optional login system for users
10. 🎨 Add dark / light mode switch
11. ⚙️ Configurable batch size from UI
12. 🌐 Add file upload progress indicator
13. 📂 Store previous analyses locally
14. 📡 WebSocket live updates (no reload needed)
15. 🪶 Responsive UI improvements for mobile
16. 📊 Add chart view for keyword frequency
17. 🧮 Implement sentiment classification heatmap
18. 🧱 Convert to Progressive Web App (PWA)
19. 🧰 Add command-line interface (CLI) tool
20. 🤝 Integrate HuggingFace local models option

---

⭐ Contribute

Pull requests are welcome! For major changes, please open an issue first to discuss what you’d like to improve.

---

📧 Author

Hossein Fallah GitHub: @hosseinfallah-h

---

✅ Summary of What This README Includes

Section	Purpose
🔹 Intro	What the app does
🔹 Features	Key highlights
🔹 Tech Stack	Components used
🔹 Setup & Run	All exact commands
🔹 Folder Structure	Visual project map
🔹 Dataset Example	Example table
🔹 Local Model	How to change the AI model
🔹 Future Updates	20 ready milestone ideas
🔹 License & Credits	Author info