📊 Employee Review Analytics Dashboard

An interactive dashboard for analyzing employee reviews with sentiment analysis, topic modeling, and AI-powered recommendations.

🌟 Features

📈 Core Analytics

• Sentiment Classification: Automatically classifies reviews as positive, negative, or neutral
• User Type Classification: Categorizes reviewers (Intern, Senior/Manager, Consultant, etc.)
• Interactive Filtering: Filter by user type, sentiment, year, and review source
• Real-time Metrics: Overview of key statistics and trends

🎯 Advanced Analysis

• Topic Modeling: Identifies key themes in reviews using LDA (Latent Dirichlet Allocation)
• Word Clouds: Visual representation of most common words in positive/negative reviews
• Trend Analysis: Sentiment trends over time
• Rating Distribution: Detailed breakdown of review ratings

🔍 Search & Browse

• Smart Search: Search reviews by keywords, job titles, or content
• Advanced Sorting: Sort by recency, rating, helpfulness
• Review Details: Expandable view with full review content and metadata

💡 AI-Powered Recommendations

• Critical Issues Analysis: Identifies most common problems from negative reviews
• Gemini AI Integration: Generates specific, actionable improvement recommendations
• Export Capabilities: Download filtered data and summary reports
• Keyword Tracking: Monitor specific terms and get alerts

📊 Visualizations

• Interactive Charts: Built with Plotly for responsive, interactive visualizations
• Sentiment Distribution: Pie charts, bar charts, and trend lines
• User Type Analysis: Breakdown by employee categories
• Rating Analytics: Comprehensive rating distribution analysis

🚀 Quick Start

Option 1: Windows Batch File (Easiest)

1. Double-click run_dashboard.bat
2. The script will automatically install dependencies and start the dashboard
3. Open your browser to http://localhost:8501

Option 2: Python Setup Script

python setup_dashboard.py

Option 3: Manual Installation

# Install dependencies
pip install -r requirements.txt

# Run the dashboard
streamlit run dashboard.py

📋 Requirements

System Requirements

• Python 3.8 or higher
• Windows, macOS, or Linux
• Internet connection (for AI recommendations)

Data Requirements

• merged_reviews.json file in the same directory
• The file should contain the merged review data from Glassdoor and Indeed

Python Dependencies

streamlit>=1.28.0
pandas>=1.5.0
plotly>=5.15.0
numpy>=1.21.0
textblob>=0.17.1
wordcloud>=1.9.2
scikit-learn>=1.3.0
nltk>=3.8
requests>=2.28.0
python-dateutil>=2.8.2

🎛️ Dashboard Usage

🔍 Filters Panel (Left Sidebar)

• User Type: Filter by employee category (Intern, Senior/Manager, etc.)
• Sentiment: Show only positive, negative, or neutral reviews
• Year Range: Select specific time periods
• Review Source: Filter by Glassdoor or Indeed
• Keyword Tracking: Monitor specific terms with real-time alerts

📊 Main Dashboard Tabs

1. Overview Tab

• Overall sentiment distribution
• Sentiment by user type
• Sentiment trends over time
• Rating distribution analysis

2. Topic Analysis Tab

• Automated topic modeling for each sentiment category
• Key themes and patterns identification
• Topic distribution visualization

3. Word Clouds Tab

• Visual representation of most frequent words
• Separate clouds for positive and negative reviews
• Interactive and responsive design

4. Review Search Tab

• Powerful search functionality
• Sort and filter capabilities
• Detailed review browsing
• Expandable review cards with full metadata

5. Recommendations Tab

• AI-powered analysis of negative reviews
• Critical issues identification
• Actionable improvement suggestions
• Export and reporting capabilities

🤖 AI Integration

Gemini API Features

• Smart Analysis: Uses Google's Gemini AI for intelligent review analysis
• Actionable Insights: Generates specific, implementable recommendations
• Priority Assessment: Categorizes issues by importance and urgency
• Success Metrics: Suggests measurable outcomes for tracking improvement

API Configuration

The dashboard uses the provided Gemini API key: AIzaSyDmPiDZGWkap0oKZzENszs9nsK-E-A2qMA

📁 File Structure

Sentiment Analysis/
├── dashboard.py              # Main dashboard application
├── setup_dashboard.py       # Python setup script
├── run_dashboard.bat        # Windows launcher
├── requirements.txt         # Python dependencies
├── README.md               # This file
├── merge_reviews.py        # Review merging script
├── merged_reviews.json     # Processed review data
├── Glassdoor.json         # Original Glassdoor data
└── Indeed.json            # Original Indeed data

🔧 Technical Details

Data Processing Pipeline

1. Data Loading: Reads merged JSON review data
2. Text Processing: Cleans and preprocesses review text
3. Sentiment Analysis: Uses VADER and TextBlob for sentiment scoring
4. Classification: Categorizes users based on job titles and employment data
5. Topic Modeling: Applies LDA for theme extraction
6. Visualization: Generates interactive charts and graphs

Machine Learning Components

• Sentiment Analysis: Hybrid approach using VADER + TextBlob
• Topic Modeling: Latent Dirichlet Allocation (LDA)
• Text Vectorization: TF-IDF for feature extraction
• Clustering: K-means for review grouping

Performance Optimization

• Caching: Streamlit session state for data persistence
• Lazy Loading: Components loaded on demand
• Efficient Processing: Vectorized operations with pandas/numpy
• Memory Management: Optimized data structures

🚨 Troubleshooting

Common Issues

"No module named 'streamlit'"

pip install streamlit

"merged_reviews.json not found"

• Ensure the data file is in the same directory as the dashboard
• Run the merge script first if needed

Dashboard won't start

• Check Python version (3.8+ required)
• Verify all dependencies are installed
• Check firewall settings for port 8501

AI recommendations not working

• Verify internet connection
• Check if API key is valid
• Review API quota limits

Performance Tips

• Use filters to reduce dataset size for faster processing
• Close other browser tabs when running the dashboard
• Restart the dashboard if it becomes slow

📈 Analytics Capabilities

Sentiment Metrics

• Overall Distribution: Percentage breakdown of sentiment categories
• Temporal Trends: How sentiment changes over time
• User Type Analysis: Sentiment patterns by employee category
• Source Comparison: Differences between Glassdoor and Indeed reviews

Topic Analysis

• Automatic Theme Detection: Identifies key discussion topics
• Sentiment-Specific Topics: Different themes for positive/negative reviews
• Word Frequency Analysis: Most common terms and phrases
• Trend Identification: Emerging issues and positive developments

Recommendation Engine

• Issue Prioritization: Ranks problems by frequency and impact
• Action Planning: Specific steps for improvement
• Success Tracking: Metrics for measuring progress
• Timeline Planning: Realistic implementation schedules

🎯 Use Cases

HR Teams

• Monitor employee satisfaction trends
• Identify systemic issues early
• Generate improvement action plans
• Track progress over time

Management

• Understand team sentiment
• Prioritize workplace improvements
• Benchmark against industry standards
• Make data-driven decisions

Executives

• High-level satisfaction overview
• Strategic planning insights
• Risk identification
• Culture assessment

🔒 Privacy & Security

• All processing is done locally
• No sensitive data is transmitted (except to Gemini for recommendations)
• Review anonymization built-in
• Secure API key handling

🤝 Contributing

Feel free to submit issues, feature requests, or pull requests to improve the dashboard.

📄 License

This project is for internal use and analysis purposes.

---

Happy Analyzing! 📊✨