BiomarkerDash

Real-time Biomarker Monitoring Dashboard

A full-stack clinical dashboard that streams patient biomarker data over WebSockets, detects anomalies with ML, and escalates clinical alerts against validated reference ranges.

Methods & contributions

• Clinically accurate ranges — reference intervals and critical thresholds verified against ICU conventions (K⁺ > 6.5, glucose < 40, SpO₂ < 88, troponin > 0.4).
• Anomaly detection — Isolation Forest plus z-score and OLS-trend analysis on streaming vitals.
• Real-time transport — WebSocket streaming with Canvas-rendered charts and Redis-backed state.
• Alert escalation — tiered clinical alerting with sound clinical-value logic.

Data flow

device stream → WebSocket → anomaly detector (IForest · z-score)
             → alert engine (critical thresholds) → live Canvas dashboard

Technologies