🎬 Movie Review Sentiment Analysis

🤖 Professional VADER Sentiment Analysis on IMDb Movie Reviews

Comprehensive NLP project analyzing 2,000 movie reviews with advanced error analysis and statistical evaluation

🎯 Project Overview

This project implements VADER (Valence Aware Dictionary and sEntiment Reasoner) sentiment analysis on the famous IMDb Movie Reviews Dataset containing 2,000 movie reviews. The goal is to automatically classify movie reviews as positive or negative sentiment and evaluate the performance against human-labeled ground truth.

🔍 Key Highlights

Dataset: 2,000 IMDb movie reviews (1,000 positive, 1,000 negative)
Algorithm: NLTK VADER Sentiment Intensity Analyzer
Performance: 63.67% accuracy with detailed analysis
Challenge: Understanding complex human semantics in movie reviews
Analysis: 20+ professional visualizations and statistical tests

🧠 What Makes This Project Special

Unlike simple text classification, this project tackles the challenging problem of sentiment analysis in movie reviews, where:

Reviews often contain mixed sentiments
Sarcasm and irony are common
Context matters more than individual words
Final judgments may appear only at the end of long reviews

📊 Key Results

✅ Overall Accuracy: 63.67%
📊 Total Reviews Analyzed: 1,938
✔️ Correct Predictions: 1,234
❌ Incorrect Predictions: 704

📈 Performance Breakdown:
• Precision (Negative): 72.0%
• Precision (Positive): 60.0%
• Recall (Negative): 44.0%
• Recall (Positive): 83.0%
• F1-Score (Macro): 62.0%

🚀 Quick Start

Prerequisites

Python 3.7+
pip install -r requirements.txt

Installation

# Clone the repository
git clone https://github.com/Ahmadhammam03/movie-sentiment-analysis.git
cd movie-sentiment-analysis

# Create virtual environment
python -m venv sentiment_env
source sentiment_env/bin/activate  # On Windows: sentiment_env\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Download NLTK data
python -c "import nltk; nltk.download('vader_lexicon')"

Run the Analysis

# Launch Jupyter Notebook
jupyter notebook

# Open and run the enhanced notebook
# -> Sentiment-Analysis.ipynb

📁 Project Structure

movie-sentiment-analysis/
│
├── 📓 Sentiment-Analysis.ipynb                       # Main analysis notebook
├── 📊 data/
│   └── moviereviews.tsv                              # Movie reviews dataset
├── 🖼️ assets/                                        # Generated visualizations
│   ├── sentiment_analysis_banner.png                 # Project banner
│   ├── confusion_matrix.png                          # Confusion matrix visualization
│   ├── sentiment_distribution.png                    # Data distribution plot
│   ├── methodology_flow.png                          # Process flow diagram
│   └── error_examples.png                            # Error analysis examples
├── 📋 requirements.txt                               # Python dependencies
├── 📄 README.md                                      # Project documentation
├── 📜 LICENSE                                        # MIT License
└── 🚫 .gitignore                                     # Git ignore rules

🔬 Methodology

1. Data Preprocessing

Data Loading: Import 2,000 movie reviews from TSV file
Data Cleaning: Remove NaN values and empty strings
Quality Assessment: Analyze text length and characteristics

2. VADER Sentiment Analysis

from nltk.sentiment.vader import SentimentIntensityAnalyzer

# Initialize VADER analyzer
sid = SentimentIntensityAnalyzer()

# Generate sentiment scores
scores = sid.polarity_scores(review_text)
# Returns: {'neg': 0.121, 'neu': 0.778, 'pos': 0.101, 'compound': -0.9125}

3. Performance Evaluation

Classification Metrics: Accuracy, Precision, Recall, F1-Score
Error Analysis: False positives/negatives with examples
Statistical Testing: Mann-Whitney U tests for significance
Visualization: 20+ professional charts and dashboards

📈 Key Findings

🎯 VADER Strengths:

✅ High recall for positive reviews (83%) - excellent at finding positive sentiment
✅ Good precision for negative reviews (72%) - confident negative predictions
✅ Fast processing (500+ reviews/second)
✅ No training required - ready-to-use approach
✅ Interpretable results with sentiment components

⚠️ VADER Limitations:

❌ 542 false positives - struggles with negative reviews containing positive words
❌ Limited context understanding beyond individual words
❌ Difficulty with sarcasm, irony, and mixed sentiments
❌ Challenges with delayed judgment patterns in movie reviews

💡 Research Insights:

Movie reviews exhibit complex sentiment structures not easily captured by lexicon-based methods
Many reviews discuss positive aspects before delivering negative final judgment
Context and discourse structure matter more than individual word sentiment
The decision boundary may need domain-specific optimization

🎨 Visualizations

The project includes comprehensive visualizations:

Dataset Analysis Dashboard (4-panel overview)
Sentiment Score Analysis (9-panel comprehensive view)
Performance Evaluation Dashboard (9-panel metrics analysis)
Error Analysis Visualizations (detailed error patterns)
Final Summary Dashboard (key findings and recommendations)

🛠️ Technical Stack

Core Libraries

pandas: Data manipulation and analysis
numpy: Numerical computing
nltk: Natural Language Processing toolkit with VADER
scikit-learn: Machine learning metrics and evaluation
matplotlib/seaborn: Professional data visualization
scipy: Statistical analysis and significance testing

📚 Dataset Information

Source

Original Dataset: Cornell Movie Review Data v2.0
Authors: Bo Pang and Lillian Lee (2004)
Publication: "A Sentimental Education: Sentiment Analysis Using Subjectivity Summarization Based on Minimum Cuts", ACL 2004
Size: 2,000 movie reviews from IMDb
Balance: Perfectly balanced (1,000 positive, 1,000 negative)

Citation

@InProceedings{Pang+Lee:04a,
  author = {Bo Pang and Lillian Lee},
  title = {A Sentimental Education: Sentiment Analysis Using Subjectivity Summarization Based on Minimum Cuts},
  booktitle = "Proceedings of the ACL",
  year = 2004
}

🔮 Future Improvements

📈 Immediate Enhancements

Adjust decision threshold based on domain-specific validation
Implement preprocessing to handle negations more effectively
Add movie-specific sentiment lexicon expansions
Consider sentence-level sentiment analysis for longer reviews

🚀 Advanced Approaches

Implement transformer-based models (BERT, RoBERTa) for better context understanding
Use ensemble methods combining VADER with other sentiment approaches
Develop domain-specific models trained on movie review corpora
Apply attention mechanisms to identify key sentiment-bearing passages

Learning Outcomes

Technical Skills Developed

Sentiment Analysis: Understanding VADER lexicon-based approach
Data Preprocessing: Cleaning and preparing text data for analysis
Performance Evaluation: Using comprehensive classification metrics
Statistical Analysis: Conducting significance tests and error analysis
Data Visualization: Creating professional, publication-ready charts

Domain Knowledge Gained

NLP Challenges: Real-world complexities in sentiment analysis
Human Language: Understanding context and semantic nuances
Model Limitations: When rule-based systems fall short
Research Methods: Systematic evaluation and error analysis

🤝 Contributing

Contributions are welcome! Here's how you can help improve this project:

🛠️ Areas for Contribution

Add additional sentiment analysis methods for comparison
Implement deep learning approaches (BERT, RoBERTa)
Create interactive web interface for real-time analysis
Add more comprehensive error analysis categories
Improve documentation and add more examples

📝 How to Contribute

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

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

👨‍💻 Author

Ahmad Hammam

GitHub: @Ahmadhammam03
LinkedIn: Ahmad Hammam

Skills Demonstrated in This Project:

Advanced Python programming and data analysis
Natural Language Processing with NLTK
Statistical analysis and hypothesis testing
Professional data visualization and storytelling
Research methodology and systematic evaluation
Technical documentation and communication

🙏 Acknowledgments

Cornell University for providing the movie review dataset
NLTK Team for the excellent VADER sentiment analysis tool
Open Source Community for making advanced NLP accessible to everyone

📊 Project Stats

⭐ If you found this project helpful, please consider giving it a star! ⭐

🎬 "In the world of sentiment analysis, context is king and VADER is a worthy knight!" 🎭

Name		Name	Last commit message	Last commit date
assets		assets
data		data
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
Sentiment-Analysis.ipynb		Sentiment-Analysis.ipynb
requirements.txt		requirements.txt

Latest commit

History

Folders and files

Repository files navigation