π§ Data Scientist β’ π€ Full-Stack Web Developer β’ π Insight Architect
Transforming complex data into actionable intelligence
|
π Academic Foundation: Information Technology @ Nelson Mandela University π Portfolio: lesleyngcobo.netlify.app - Explore my complete project showcase |
| π― DataQuest Recommender | π‘οΈ Fraud Detection System | π₯ Madiba HealthConnect | π Portfolio Website |
|---|---|---|---|
| Hybrid ML Recommender | Advanced Fraud Detection | Healthcare Management Platform | Client Portfolio Site |
| Collaborative filtering + content-based algorithms for personalized recommendations | Imbalanced classification with ensemble learning and real-time monitoring | Full-stack digital health app for patient records, appointments, and clinician workflows | Modern responsive design with SEO optimization |
R recommenderlab Machine Learning |
R Power BI XGBoost SMOTE |
ASP.NET Core 8 Angular Node.js |
HTML5 CSS3 JavaScript |
| View Project β | View Project β | View Project β β’ Live Demo β | View Live β |
-
DataQuest Recommender:
π Item-Based Collaborative Filtering (IBCF) outperformed all models with 22Γ higher precision and 15Γ higher recall than UBCF.
π Final deployed Full Hybrid balanced accuracy with contextual signals (region + season) for real-world adaptability. -
Fraud Detection System:
β‘ XGBoost achieved 79% recall and 94% precision, minimizing false alarms while catching most fraud cases.
π Improved recall by ~17% compared to baseline logistic regression, making fraud detection both accurate and practical.
|
R & RStudio |
![]() Python |
![]() Tidyverse |
![]() ggplot2 |
recommenderlab |
|
Power BI |
Excel |
![]() MySQL |
SQL Server |
SSMS |
|
Angular |
HTML5 |
CSS3 |
JavaScript |
ASP.NET |
graph TD
A[π§ Machine Learning Mastery] --> B[π Hybrid Recommender Systems]
A --> C[π‘οΈ Fraud Detection Models]
A --> D[π Ensemble Methods]
B --> E[Collaborative Filtering]
B --> F[Content-Based Systems]
C --> G[Imbalanced Classification]
C --> H[Real-time Monitoring]
D --> I[Random Forests]
D --> J[XGBoost Optimization]
style A fill:#64ffda,stroke:#0f0f23,color:#0f0f23
style B fill:#4ecdc4,stroke:#0f0f23,color:#0f0f23
style C fill:#ff6b6b,stroke:#0f0f23,color:#0f0f23
style D fill:#ffd93d,stroke:#0f0f23,color:#0f0f23
π Ready to collaborate? Let's turn insights into action with data-driven intelligence!
π¬ Get in touch: Portfolio β’ Email β’ LinkedIn
β‘ Made with β€οΈ and lots of β by Lesley Ngcobo
Last updated: August 2025 β’ Data Scientist & ML Specialist






