How phil27-gotchu/jaundice_detection works
Jaundice Monitor combines an ESP8266 reflectance probe with a dashboard/API, plus a Cloudflare deployment variant and logistic-regression screening. The main documented workflow is probe measurement and local logging, then HTTP upload, server-side reading handling, and browser dashboard rendering; inference enriches displayed readings. Python training and prediction modules form a separate model-development branch. Some links are documented rather than visible in sampled firmware, and the SQLite helper’s runtime use is not established by the excerpts.
Main entry points: jaundice_monitor.ino Probe Firmware · index.js Local API · index.ts Cloudflare API · dashboard.js Dashboard Renderer
Components 19
Probe Firmware
- Probe Firmware
firmware/jaundice_monitor.ino
Ingest Services
- Local API
server/index.js - Cloudflare API
cloudflare/src/index.ts - Readings KV
Dashboard
- Dashboard Template
cloudflare/public/index.html - Dashboard Styles
cloudflare/public/styles.css - Dashboard Renderer
shared/dashboard.js
Screening Model
- JS Predictor
shared/mlPredictor.js - Trained Model
ml/model/trained_model.json
Model Development
- Synthetic Dataset
ml/data/synthetic_neonatal_jaundice_dataset.csv - Model Training
ml/model/train_and_export.py - Python Predictor
ml/model/predict.py - Logistic Regression
ml/model/logistic_regression.py - Random Forest
ml/model/random_forest.py
Other components
- Operator
- RGB Probe
- OLED Display
- SD Card Log
- Browser
Connections 15
- Operator → Probe Firmware: presses button
README.md - Probe Firmware → RGB Probe: measures RGB
README.md - Probe Firmware → OLED Display: shows results
README.md - Probe Firmware → SD Card Log: appends readings
README.md - Probe Firmware → Local API: POSTs readings
README.md - Browser → Local API: polls dashboard
README.md - Local API → Dashboard Renderer: renders page
server/index.js - Local API → JS Predictor: predicts risk
server/index.js - Cloudflare API → Dashboard Renderer: renders page
cloudflare/src/index.ts - Cloudflare API → Readings KV: reads and writes
cloudflare/src/index.ts - Cloudflare API → Dashboard Template: fetches template
cloudflare/src/index.ts - Dashboard Renderer → JS Predictor: predicts readings
shared/dashboard.js - Model Training → Synthetic Dataset: loads data
ml/model/train_and_export.py - Model Training → Trained Model: exports parameters
ml/model/train_and_export.py - JS Predictor → Trained Model: loads model
shared/mlPredictor.js