---
title: Advanced Sentiment Analytics Dashboard
emoji: 📊
colorFrom: blue
colorTo: green
sdk: docker
pinned: false
license: mit
app_port: 7860
---

# 🚀 Advanced Sentiment Analytics Dashboard

A distributed sentiment analysis system with **PyABSA + HF Spaces backend** and **Streamlit Cloud frontend** for scalable, high-accuracy analysis of product reviews.

## 🏗️ Architecture

### 🔧 Split Deployment Strategy
- **Backend**: HuggingFace Spaces (PyABSA + FastAPI) - `api_server.py`
- **Frontend**: Streamlit Cloud (Lightweight UI) - `frontend_light.py`
- **Benefits**: High accuracy models + fast, scalable frontend

## ✨ Features

### 🔍 Core Analytics
- **PyABSA Integration**: State-of-the-art aspect-based sentiment analysis
- **M2M100 Translation**: Facebook's multilingual translation model
- **Intent Classification**: Advanced customer intent detection
- **Real-time Processing**: API-based architecture for scalability

### 📊 Advanced Dashboard
- **Multi-page Navigation**: Home, Analytics, History, Documentation
- **Interactive Visualizations**: Network graphs, heatmaps, timelines
- **Advanced Filtering**: Multi-dimensional data exploration
- **Session Management**: Save and restore analysis sessions
- **KPI Dashboard**: Real-time metrics and insights

### 🎯 Business Intelligence
- **Areas of Improvement**: AI-powered identification of problem areas
- **Strength Anchors**: Recognition of positive aspects to leverage
- **Alert System**: Automated sentiment spike detection
- **Impact Simulation**: What-if analysis for business decisions
- **Export Functionality**: PDF reports and Excel data export

## 📝 Data Format

Your CSV file should include these columns:
- `id`: Unique identifier for each review
- `reviews_title`: Title of the review
- `review`: The actual review text
- `date`: Review date (YYYY-MM-DD format)
- `user_id`: Identifier for the reviewer

## 🚀 Usage

### For End Users:
1. **Upload your CSV file** with review data (see format below)
2. **Process the data** using our advanced NLP pipeline (~2-3 minutes for 1000 reviews)
3. **Explore insights** through interactive visualizations
4. **Export results** as PDF reports or Excel files

### For Developers:
#### Docker Deployment (Hugging Face Spaces)
This app is optimized for Docker deployment with:
- `Dockerfile`: Production-ready container setup
- `requirements-docker.txt`: Optimized dependencies with version pinning
- `.dockerignore`: Efficient build context
- Health checks and proper port configuration (7860)

#### Local Development
```bash
git clone <your-repo>
cd insights
pip install -r requirements.txt
streamlit run app_enhanced.py
```

## 🛠️ Technology Stack

- **Frontend**: Streamlit with interactive components
- **NLP**: pyABSA for aspect-based sentiment analysis
- **Translation**: Facebook M2M100 for multilingual support
- **Visualization**: Plotly for interactive charts and graphs
- **Network Analysis**: NetworkX for aspect relationship graphs

## 📊 Sample Output

The dashboard provides:
- Comprehensive sentiment analysis
- Aspect extraction and sentiment mapping
- Intent classification with confidence scores
- Interactive network graphs of aspect relationships
- Time-series analysis of sentiment trends
- Exportable business intelligence reports

---

**Status**: ✅ **Production Ready** - Enterprise-level sentiment analysis with advanced NLP capabilities.