🔬 Research Analyst · 📊 Data Scientist · 📈 Statistician
Research analyst and data scientist with expertise in statistical modelling, ML pipelines, and Monitoring, Evaluation, and Learning (MEL) frameworks for evidence-based decision making.
Translating complex data into actionable insights for policy, public health, and operational decision-making.
Regression pipeline predicting airline delay minutes from flight volume and carrier data.
| Metric | Result |
|---|---|
| R² Score | ~0.95 on monthly aggregates |
| Key Insight | Carrier identity explains ~30% of delay variance beyond traffic volume |
| Stack | Python · Pandas · Scikit-learn · Seaborn · Jupyter |
End-to-end ML workflow for diabetes risk prediction using the Pima Indians dataset.
| Metric | Result |
|---|---|
| Accuracy | ~75% with Logistic Regression |
| Segmentation | K-Means (k=3) identifies distinct metabolic-risk patient profiles |
| Stack | Python · Pandas · Scikit-learn · Matplotlib |
Compartmental ODE model evaluating educational campaign impact on alcohol-use population dynamics.
| Metric | Result |
|---|---|
| Model Type | SIR-style compartmental ODE |
| Output | Equilibrium behaviour & sensitivity analysis via phase-plane visualisation |
| Stack | Python · SciPy · NumPy · Matplotlib |
Statistical and network-based analysis of crime patterns for intelligence-led decision making.
| Metric | Result |
|---|---|
| Method | Network centrality & exploratory spatial analysis |
| Insight | Identifies core distribution nodes and brokerage roles in urban networks |
| Stack | Python · NetworkX · Pandas · Matplotlib |
All featured repositories are built to these standards:
- ✅ Reproducible — Every repo includes
requirements.txtand step-by-step execution instructions - ✅ Documented — Methodologies, assumptions, and limitations are explicitly stated
- ✅ Licensed — Open-source licenses for reuse and collaboration
- ✅ Decision-focused — Outputs designed for policy-makers and operational teams, not just technical peers
- 📊 Statistical Modelling — Inference, regression, and experimental design
- 🤖 Machine Learning — End-to-end pipelines for regression, classification, and clustering
- 🏥 Public Health Research — Epidemiological analysis and intervention modelling
- 🗺️ Crime & Geospatial Analysis — Pattern detection, network analysis, and spatial statistics
- 📋 MEL Systems — Monitoring, Evaluation, and Learning framework design and implementation
- 🔬 Mathematical Modelling — Compartmental ODEs for assessing public health intervention impact
| Category | Tools |
|---|---|
| Languages | Python · R · SQL · TypeScript |
| Data & ML | Pandas · Scikit-learn · Statsmodels · NumPy |
| Visualization | Matplotlib · Seaborn · Power BI · ggplot2 |
| Data Collection | KoboToolbox · REDCap · ODK · SPSS · STATA |
| Workflows | Git · Jupyter · RMarkdown · GitHub Actions |
Based in Nairobi, Kenya 🇰🇪
Work focuses on applying statistical and machine learning methods to public health, crime analysis, and decision-support systems.
Principles: Reproducibility · Interpretability · Policy-relevant insights
Contributions and replications are welcome. All flagship repositories include open-source licenses and end-to-end execution instructions.
- 📚 Developed instructional materials for undergraduate mathematics (Algebra, Calculus)
- 📝 Produced LaTeX-based academic and technical documents
- 🎓 Supported training in data collection tools (KoboToolbox, ODK) and reproducible analysis workflows
- 🗺️ Crime pattern analysis using geospatial and statistical methods
- 🔁 Strengthening reproducible research workflows with CI/CD for Jupyter notebooks
- 📊 Building MEL dashboards for public health programme evaluation
- Research collaborations in public health, epidemiology, and social policy
- MEL system design, implementation, and third-party evaluation
- Data science and analytics consultancy for NGOs and government agencies
- Academic partnerships, peer review, and teaching opportunities
Open to research collaborations, MEL consultancy, and data-driven projects.