An advanced, AI-powered medical diagnostic tool built for modern healthcare applications. This system predicts breast tumor malignancy using the Wisconsin Breast Cancer Dataset, featuring Explainable AI (XAI), Population Analytics, AI-powered recommendations, and comprehensive reporting capabilities.
This application demonstrates a complete machine learning pipeline for medical diagnostics, incorporating best practices in software engineering, data science, and user experience design. The system provides real-time predictions with transparency and actionable insights for healthcare professionals.
- Real-time Prediction: Instant benign/malignant classification with confidence scores
- Intuitive Interface: Modern, clean UI with numeric input grids (1-10 scale)
- High Accuracy: Machine learning model trained on Wisconsin Breast Cancer Dataset
- Batch Processing: Support for multiple patient records via predictive system
- Population Comparison: Interactive radar charts comparing patient biomarkers against population averages
- Outlier Detection: Visual identification of abnormal features
- Statistical Insights: Comprehensive analysis of patient data relative to historical trends
- Interactive Visualizations: Powered by Plotly for dynamic data exploration
- Feature Importance: Bar charts showing which biomarkers contributed to the prediction
- Decision Transparency: Clear visualization of model reasoning
- Confidence Metrics: Detailed probability scores for each prediction
- Interpretable Results: Making black-box models understandable for clinicians
- Google Gemini Integration: Personalized medical advice based on patient data
- Contextual Insights: Recommendations tailored to specific biomarker patterns
- Next Steps Guidance: Suggested tests and treatment considerations
- Evidence-based Responses: AI-generated advice grounded in medical context
- Feedback Collection: Clinician correction logging system
- Model Maintenance: Data collection for future retraining
- Quality Assurance: Built-in mechanisms for prediction validation
- Audit Trail: Comprehensive logging of predictions and corrections
- PDF Generation: One-click download of comprehensive medical reports
- Complete Documentation: Includes all inputs, predictions, visualizations, and recommendations
- Shareable Format: Professional reports for medical records and consultations
- Algorithm: Support Vector Machine (SVM) / Logistic Regression
- Training Data: Wisconsin Breast Cancer Dataset (UCI ML Repository)
- Features: 10 quantitative biomarkers from fine needle aspirate (FNA) imaging
- Performance: Optimized for medical diagnostic accuracy
- Backend: Python 3.8+
- Web Framework: Streamlit
- ML Libraries: scikit-learn, NumPy, pandas
- Visualization: Matplotlib, Plotly
- AI Integration: Google Generative AI (Gemini)
- PDF Generation: FPDF
- Environment Management: python-dotenv
- Python 3.8 or higher
- pip package manager
- Google Gemini API key (for AI recommendations feature)
-
Clone the Repository
git clone https://github.com/abdull6771/Breast-Cancer-Wisconsin-Classification.git cd Breast-Cancer-Wisconsin-Classification -
Create Virtual Environment (Recommended)
python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate
-
Install Dependencies
pip install -r requirements.txt
-
Configure Environment Variables
Create a
.envfile in the project root:GEMINI_API_KEY=your_google_gemini_api_key_here
Note: Get your API key from Google AI Studio
Start the Streamlit server:
streamlit run app.pyThe application will open in your browser at http://localhost:8501
For batch predictions or integration with other systems:
python predictive_system.pyThe system accepts the following biomarkers (scale 1-10):
| Feature | Description |
|---|---|
| Clump Thickness | Thickness of cell clumps |
| Uniformity of Cell Size | Consistency in cell dimensions |
| Uniformity of Cell Shape | Consistency in cell morphology |
| Marginal Adhesion | Cell-to-cell adhesion quality |
| Single Epithelial Cell Size | Size of individual epithelial cells |
| Bare Nuclei | Nuclei not surrounded by cytoplasm |
| Bland Chromatin | Chromatin texture uniformity |
| Normal Nucleoli | Nucleoli appearance |
| Mitoses | Rate of cell division |
Breast-Cancer-Wisconsin-Classification/
│
├── app.py # Main Streamlit application
├── predictive_system.py # Batch prediction script
├── report_generator.py # PDF report generation module
├── trained_model.sav # Trained ML model (pickle)
├── breast-cancer-wisconsin.csv # Dataset for analytics
├── requirements.txt # Python dependencies
├── README.md # Project documentation
├── .env # Environment variables (not in repo)
└── __pycache__/ # Python cache files
Wisconsin Breast Cancer Dataset (Original)
- Source: UCI Machine Learning Repository
- Instances: 699 samples
- Features: 10 quantitative attributes
- Classes: Benign (2) / Malignant (4)
- Missing Values: Handled during preprocessing
- Citation: Wolberg, W.H., & Mangasarian, O.L. (1990)
All features are scored on a 1-10 scale based on FNA imaging analysis:
- Values 1-3: Generally benign characteristics
- Values 4-6: Intermediate patterns
- Values 7-10: Typically malignant indicators
Clean, intuitive input grid with real-time prediction
Radar chart comparison against benign/malignant averages
Feature contribution visualization showing model reasoning
Personalized medical advice from Google Gemini
(Add screenshots to your repository and link them here)
- 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
Run the application in development mode:
streamlit run app.py --server.runOnSave trueThe trained model demonstrates strong performance on the Wisconsin dataset:
- High accuracy in binary classification
- Robust feature importance analysis
- Validated against clinical standards
(Update with specific metrics if available)
Contributions are welcome! Please feel free to submit a Pull Request. For major changes, please open an issue first to discuss what you would like to change.
This project is licensed under the MIT License - see the LICENSE file for details.
IMPORTANT: This application is designed for educational and research purposes only. It is NOT intended to replace professional medical diagnosis, advice, or treatment. Always consult qualified healthcare professionals for medical decisions.
The predictions made by this system should be used as supplementary information only and must be validated by licensed medical practitioners.
- UCI Machine Learning Repository for the Wisconsin Breast Cancer Dataset
- Streamlit for the amazing web framework
- Google AI for Gemini API access
- scikit-learn community for ML tools
- The open-source community for continuous inspiration
Abdullah
- GitHub: @abdull6771
- Project Link: https://github.com/abdull6771/Breast-Cancer-Wisconsin-Classification
- Add support for additional ML algorithms
- Implement model versioning and A/B testing
- Add user authentication and role-based access
- Expand dataset with more recent clinical data
- Create mobile-responsive design
- Add multi-language support
- Integrate with Electronic Health Records (EHR) systems
- Deploy to cloud platforms (AWS/Azure/GCP)
Made with ❤️ for advancing healthcare through AI
⭐ Star this repository if you find it helpful!