-
Notifications
You must be signed in to change notification settings - Fork 0
Home
Batuhan4 edited this page Apr 20, 2026
·
10 revisions
ML Visualization Tool for Healthcare Professionals SENG 430 · Software Quality Assurance · Çankaya University · Spring 2025-2026
Welcome to the project wiki! This is the central hub for all documentation, meeting notes, architecture decisions, and sprint retrospectives.
| Page | Description |
|---|---|
| Home | This page — project overview and navigation |
| Team | Team members, roles, and contact info |
| Architecture | System architecture, service map, tech decisions |
| API | REST API reference — endpoints, request/response schemas, error codes |
| Meeting Notes | Weekly meeting notes and stand-up summaries |
| Sprint 1 | Sprint 1 — backlog, goals, deliverables, retrospective |
| Sprint 2 | Sprint 2 — Steps 1–3, QA reports, metrics, retrospective |
| Sprint 3 | Sprint 3 — Steps 4–5, 8 ML models, QA (46 TC), retrospective |
| Sprint 4 | Sprint 4 — Steps 6–7, Explainability, Ethics & Bias, PDF Certificate, retrospective |
| Sprint 5 | Sprint 5 — Polish, Lighthouse 93/100, Docker, Gemma 4 insights, user testing, jury |
| Final Submission | Week 11 jury showcase — checklist, live surfaces, 10-min deck outline, FE prep |
| Accessibility Log | Sprint 5 WCAG 2.1 AA fix log — 91 → 100 with before/after diffs |
| Domain Clinical Review | Clinical justification table for all 20 medical specialties |
| Clinical Tooltip Review | Sprint 3 deliverable — clinical tooltips for all 8 model parameter panels |
An interactive, browser-based tool that guides healthcare professionals through a 7-step ML pipeline:
- Clinical Context — Choose a medical specialty (20 domains)
- Data Exploration — Upload CSV or use built-in datasets
- Data Preparation — Handle missing values, normalise, split
- Model & Parameters — Select from 8 ML models, tune via sliders
- Results — Performance metrics, confusion matrix, ROC curve
- Explainability — Feature importance + SHAP explanations
- Ethics & Bias — Subgroup fairness audit + EU AI Act checklist
- Frontend: React 18 + Vite + TypeScript
- Backend: FastAPI (Python 3.12)
- ML Engine: scikit-learn + XGBoost + LightGBM (8 classifiers)
- Explainability: SHAP
- PM: Jira
- Design: Figma