Projects

Selected work.

A short, honest list. Two projects, both finished, both documented end to end — including the parts that didn't work.

2 projects

  1. HF-RISK — Predicting Heart Failure Outcomes with Machine Learning

    Leakage-clean ML pipeline on 2,008 patients: five models compared, 143 final predictors, SHAP-driven explainability.

    • MSc Dissertation
    • Random Forest
    • XGBoost
    • SHAP
    • MIMIC-IV

    2026

    London Met

    Clinical ML

    Primary outcome: 6-month mortality at 0.710 AUROC on the held-out Zhang cohort and 0.599 on MIMIC-IV without retraining. The 2026 extension retrained inside MIMIC-IV with a patient-grouped split: GCS components added +0.0138 AUROC at 28 days and +0.0104 beyond an ICU-stay control at six months.

    Read

  2. Interactive Card Sorting — Spatial Organization & Cognitive Strategy Analysis

    How people arrange cards on an 8×8 board — across 845 recorded trials, with an interactive viewer for retry tracing.

    • Behavioural
    • Python
    • Interactive viewer

    2026

    London Met

    Behavioural

    Organised layouts correlated with completion; strategy quality added explanatory power beyond time-on-task. Live explorer for retry tracing.

    Open

Why so few?

Because a portfolio of ten unfinished experiments says more about enthusiasm than judgement. Each project here shipped with a documented method, a checked pipeline, and a written account of its limits.

More is coming: the MIMIC-IV transport and GCS work is being written up as a manuscript, the card-sorting study is heading for its own write-up, plus the day-job analytics that pays the rent. I'd rather publish four honest pieces than forty polished claims.