General Dentist building expertise in Clinical Data Science, Dental AI, and Digital Health.
I combine clinical experience in dentistry with Python, SQL, machine learning, and data visualization to build reproducible projects focused on healthcare operations and evidence-based clinical data science.
An end-to-end machine-learning portfolio project using fully synthetic, longitudinal dental appointment data to study no-show prediction under strict prediction-time and temporal-evaluation constraints.
Key components:
- Leakage-safe historical and current-appointment feature engineering
- Chronological model development and protected future evaluation
- Logistic Regression and tree-based model comparison
- Probability calibration and operational cost/capacity analysis
- Reproducible model, diagnostic, and reporting artifacts
- Streamlit transparent model-evaluation dashboard
- Explicit limitations and non-clinical-use boundaries
- Published Version 2.0.0 release
An end-to-end analytics portfolio project demonstrating how fully synthetic appointment, treatment, procedure, and payment data can support dental clinic management.
Key components:
- Reproducible generation of nine synthetic source datasets
- Normalized SQL Server relational model
- Ten documented operational and financial SQL analyses
- Python exploratory analysis and validated analytical datasets
- Three-page Power BI operations report
- Management findings, limitations, and recommended actions
- Published Version 1.0.0 release
My current portfolio combines:
- Dental operations analytics
- Clinical and operational machine learning
- Leakage-aware temporal evaluation
- Probability calibration and decision-oriented model assessment
- Reproducible analytical workflows
- Power BI and Streamlit communication layers
The next portfolio project is being defined to extend this work without duplicating the scope of the completed analytics and no-show prediction projects.
- Python, pandas, NumPy, matplotlib, scikit-learn
- SQL and SQL Server
- Power BI and DAX
- Streamlit
- Git and GitHub
- Use reproducible and clearly documented workflows
- Keep clinical claims proportional to the available evidence
- Separate descriptive associations from causal conclusions
- Prevent temporal and target leakage in predictive workflows
- Use synthetic or appropriately governed data
- Explain model limitations and the practical consequences of errors