ESP32 Link
DISCONNECTEDSensor Data
NOT RECEIVING- Transport
- USB Serial
- Baud
- 115,200
- Packets
- 0
- Last pkt
- never
Heart disease risk analysis
Heart disease risk estimation
Computed live from real ECG, PPG, SpO₂, LM35 temperature and BMI323 motion data only.
ESP32 disconnected — no physiological data to analyse.
Extracted signal features (real measurements)
- Analysis window
- 0.0 s
- ECG sampling
- --
- Detected beats
- 0
- ECG heart rate
- --
- SDNN
- --
- RMSSD
- --
- pNN50
- --
- RR irregularity
- --
- QRS amplitude
- --
- QRS duration
- --
- ST deviation
- --
- Mean SpO₂
- --
- Mean BPM (firmware)
- --
- Mean temperature
- --
- Motion (|a|)
- --
- Subject state
- --
How this estimate is produced
- R peaks are detected with a Pan-Tompkins style pipeline (derivative → squaring → 120 ms integration → adaptive threshold) on the raw AD8232 stream.
- Each condition uses an interpretable logistic model over those features. Weights come from published clinical thresholds — ST deviation for ischemia, QRS low voltage for amyloidosis, QRS widening and RR irregularity for fibrosis, resting tachycardia with desaturation for heart failure.
- A feature that has not been measured is reported as missing and simply left out of the score — it is never replaced by an assumed value.
- Single-lead ECG plus PPG cannot diagnose these diseases. Treat every percentage as a screening signal for follow-up with a clinician, not a diagnosis.
Research & monitoring use only. Sensor measurements may require calibration and clinical validation and should not be used as a substitute for professional medical diagnosis or treatment.