Customer Review Analyzer Dashboard

An AI-powered customer review analysis tool that uses state-of-the-art NLP models to analyze sentiment, categorize feedback, and extract actionable insights from customer reviews.

Link: https://review-analyzer-ai.streamlit.app/

Features

AI-Powered Analysis

• Sentiment Analysis: Classify reviews as Positive, Negative, or Neutral using BERTweet
• Category Classification: Automatically categorize reviews into topics (Product Quality, Shipping, Customer Service, etc.)
• Review Summarization: Generate concise summaries from multiple reviews using BART-CNN
• Keyword Extraction: Identify key terms from positive and negative reviews

Interactive Dashboard

• Real-time Analytics: Visualize sentiment distribution, category breakdowns, and trends
• Advanced Filtering: Filter reviews by sentiment, category, rating, and keywords
• Custom Visualizations: Interactive Plotly charts with multiple themes
• Confidence Scores: View model confidence for each prediction

Deep Dive Analysis

• Keyword Analysis: Discover most common words in positive/negative reviews
• Sentiment by Category: See how sentiment varies across different topics
• Rating Correlation: Analyze relationship between ratings and sentiment
• Review Explorer: Search and browse individual reviews with full metadata

Export Capabilities

• Download filtered results as CSV
• Export complete analyzed dataset
• Timestamped file naming for version control

Quick Start

Prerequisites

• Python 3.8 or higher
• 4GB+ RAM recommended
• GPU optional (will use CPU if not available)

Installation

1. Clone the repository

git clone https://github.com/yourusername/review-analyzer.git
cd review-analyzer

2. Create a virtual environment (recommended)

python -m venv venv

# On Windows
venv\Scripts\activate

# On macOS/Linux
source venv/bin/activate

3. Install dependencies

pip install -r requirements.txt

4. Download spaCy language model

python -m spacy download en_core_web_sm

Running the Application

streamlit run app.py

The application will open in your default browser at http://localhost:8501

Project Structure

review-analyzer/
│
├── app.py                      # Main Streamlit application
├── review_analyzer.py          # Core NLP analysis engine
├── requirements.txt            # Python dependencies
├── README.md                   # This file

Input Data Format

Your CSV file should contain review text in one of these column names:

• review_text, text, review, content, reviewText, Text, Review, reviews, or customer_review

Optional rating column (automatically detected):

• rating, star_rating, stars, score, Rating, or Star

Example CSV Format

review_text,rating
"Great product! Exceeded my expectations.",5
"Shipping was delayed by 3 days.",2
"Good value for money. Would recommend.",4
"Poor quality. Broke after one week.",1
"Decent product but overpriced.",3

Usage Guide

1. Upload & Analyze Tab

1. Upload your CSV file

   • Click "Browse files" and select your review data
   • Supported format: CSV with text and optional rating columns

2. Configure settings (in sidebar)

   • Set sample size (0 = analyze all reviews)
   • Choose visualization theme

3. Load models

   • Click "Load & Initialize Models"
   • Wait for models to download and initialize (first run takes longer)

4. Analyze reviews

   • Click "Analyze Reviews"
   • Processing time depends on dataset size (~1-2 reviews/second on CPU)

2. Dashboard Tab

View high-level analytics:

• Key Metrics: Total reviews, sentiment breakdown, confidence scores
• Sentiment Distribution: Visual breakdown of positive/negative/neutral
• Category Analysis: Most common review topics
• Confidence Distribution: Model prediction reliability
• Rating Analysis: If ratings are available

3. Deep Dive Tab

Explore detailed insights:

• Filter reviews by sentiment, category, and rating
• View top keywords for positive and negative reviews
• Generate AI summaries for specific review segments
• Analyze sentiment patterns across categories

4. Review Explorer Tab

Browse individual reviews:

• Search reviews using keywords
• Filter by multiple criteria
• Sort by confidence, rating, or other fields
• Export filtered results to CSV

Configuration

Model Settings

The application uses three transformer models:

Task	Model	Purpose
Sentiment Analysis	finiteautomata/bertweet-base-sentiment-analysis	Classify positive/negative/neutral
Category Classification	facebook/bart-large-mnli	Zero-shot topic categorization
Summarization	facebook/bart-large-cnn	Generate review summaries

Categories

Default review categories (customizable in review_analyzer.py):

• Product Quality (Defect or Condition)
• Shipping and Delivery Experience
• Customer Service and Support
• Value for Money and Pricing
• Product Features and Functionality
• Durability and Longevity

Performance Tuning

Adjust these constants in ReviewAnalyzer class:

BATCH_SIZE = 32  # Increase for faster processing (requires more RAM)
SENTIMENT_MAX_LENGTH = 128  # Maximum tokens for sentiment analysis
CLASSIFICATION_MAX_LENGTH = 512  # Maximum tokens for classification

Performance

Processing Speed

Hardware	Reviews/Second	100 Reviews	1000 Reviews
CPU (i7)	~1-2	~1 min	~10 min
GPU (RTX 3060)	~10-15	~10 sec	~2 min

Memory Usage

• Minimum: 4GB RAM
• Recommended: 8GB RAM
• With GPU: 6GB VRAM for optimal performance

Troubleshooting

Common Issues

Models not loading

# Clear transformers cache and reinstall
rm -rf ~/.cache/huggingface/
pip install --upgrade transformers

Out of memory errors

• Reduce BATCH_SIZE in review_analyzer.py
• Use sample size to analyze fewer reviews
• Close other applications

spaCy model not found

python -m spacy download en_core_web_sm

Slow performance

• Enable GPU if available
• Increase BATCH_SIZE for faster processing
• Use sample size for large datasets

Dependencies

streamlit>=1.28.0
pandas>=1.5.0
plotly>=5.17.0
transformers>=4.30.0
torch>=2.0.0
spacy>=3.6.0
numpy>=1.24.0

See requirements.txt for complete list with version pinning.

Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

Development Setup

# Install development dependencies
pip install -r requirements-dev.txt

# Run tests
pytest tests/

# Format code
black .
flake8 .

Acknowledgments

• Hugging Face for transformer models and the Transformers library
• Streamlit for the amazing web framework
• Plotly for interactive visualizations
• spaCy for NLP preprocessing

Roadmaps

Version History

v1.0.0 (Current)

• Initial release
• Sentiment analysis, categorization, summarization
• Interactive Streamlit dashboard
• Export capabilities

Disclaimer

This tool uses AI models for analysis. While generally accurate, predictions should be validated for critical business decisions. Model performance may vary based on:

• Review language and quality
• Domain-specific terminology
• Dataset size and diversity

Star History

If you find this project useful, please consider giving it a star!

---

Made with Python, Streamlit, and Transformers