Solving real-world problems — full-stack apps, data pipelines, and experiments.
Next.js
TypeScript
Supabase
Azure OpenAI
Tailwind CSS
PostgreSQL
AI-powered tutor support platform that helps tutors generate personalised lesson plans, targeted feedback, and organise student learning more efficiently. Combines full-stack development, database design, authentication, and AI integration to solve a real teaching workflow problem.
view_live()
Student Performance Predictor
Python
scikit-learn
XGBoost
Pandas
Jupyter
Multi-class classification pipeline predicting a UK university student's likely outcome — Distinction, Pass, Fail, or Withdrawn — using only their first 4 weeks of behavioural data, framed as an early-warning system for at-risk students. Compares Logistic Regression, Random Forest, and XGBoost, with feature importance analysis and a responsible-AI review that deliberately excludes protected characteristics from the model.
view_source()
Global CO₂ Emissions Analysis
Python
Pandas
Matplotlib
Seaborn
Jupyter
Exploration of the Our World in Data CO₂ dataset, visualising long-term emission trends across major countries, top emitters by total and per-capita output, the GDP–emissions relationship, and breakdowns by fuel type — coal, oil, and gas.
view_source()
