This tool helps businesses analyze their reviews by providing:
- Sentiment analysis of reviews
- Weekly summary reports
- Real-time complaint detection
- Sentiment trend visualization
- Sentiment Analysis: Uses Hugging Face's DistilBERT model for accurate sentiment classification
- Complaint Detection: Identifies negative reviews and complaints using keyword matching
- Weekly Reports: Generates PDF reports with sentiment trends and complaint summaries
- Vector Storage: Uses ChromaDB for efficient review storage and retrieval
- RAG Workflow: Implements Retrieval-Augmented Generation for better analysis
- Run the setup script:
python setup.py
This will:
- Create a virtual environment (if not exists)
- Install all required dependencies
- Download necessary models (spaCy, sentence transformers, sentiment analysis)
- Run the sample analyzer:
python sample_review_analyzer.py
This will:
- Process sample reviews from
sample_reviews.json
- Generate a weekly report
- Show sentiment analysis results
- Display detected complaints
- Create visualizations
- Weekly PDF reports in
weekly_report.pdf
- Sentiment trend visualization in
sentiment_trend.png
- Console output with analysis results
You can customize:
- Add more reviews to
sample_reviews.json
- Modify complaint keywords in
review_analyzer.py
- Adjust report format in
review_analyzer.py
When you have access to the Google Reviews API:
- Create a
.env
file with your API key:
GOOGLE_PLACES_API_KEY=your_api_key_here
- Update the
place_id
ingoogle_reviews_fetcher.py
- Run
python google_reviews_fetcher.py
This tool uses free and open-source models and libraries. No paid API keys are required for the sample data version.