The platform
Stridalyzer is the long-running web home of the smart-insole product family: clinicians and coaches review runners' activities, see foot-stress heat maps across eight plantar zones, read gait-cycle metrics (ground contact time, strike pattern, overpronation), and export branded PDF/HTML reports.
What I built
- Browser pose detection — the video gait-analysis pipeline running TensorFlow.js / MoveNet on the WebGL backend, bringing camera-based analysis into the platform with no install.
- Reporting & platform engineering — clinical report generation (PDF/HTML), charting, and feature work across the portal, shipping against live clinical users.
The platform is a large team effort — the C++→WASM analytics engine and the Three.js insole builder are colleagues' work.
How the system works
- Captureinsole sensors · uploaded video
- CloudParse activity records + cloud code
- AnalyseWASM biomechanics · TF.js pose
- Visualiseheat maps, gait charts
- Reportbranded PDF / HTML
Stack
JavaScriptTensorFlow.js / MoveNetWebGL Plotly / D3jsPDFPHPParse i18n × 6 languages