Case Study · Infant Healthcare
NOUR
ISH.
RT
Monitoring
AI
Narratives
SVG
Charts
Mobile
Native
Infant monitoring app for at-risk babies. Custom SVG growth charts, personalized baselines, and AI-powered clinical narratives via the Claude API — built solo with Claude Code.
screenshot · 1200×600
The Problem
Parents fly blind.
Parents of at-risk infants — premature babies, low birth weight, failure to thrive — are sent home with no tools to track what matters. Pediatricians get snapshots at office visits. Between appointments, parents are left guessing whether feeding patterns, weight trends, and vitals are normal or concerning.
screenshot · dashboard
What I Built
Four core systems.
01
Custom SVG Growth Charts
Hand-built SVG charts plotting weight, length, and head circumference against WHO/CDC percentile curves. No charting library — full control over rendering, annotations, and responsive behavior. Pinch-to-zoom on mobile.
screenshot · mobile
02
Personalized Baselines
Not every baby follows the 50th percentile. The app calculates individual baselines from the first 2 weeks of data, then flags deviations from that baby's trend — not a population average. Alerts are relative, not absolute.
screenshot · mobile
03
AI Clinical Narratives
Feeding logs, weight entries, and vitals are summarized into plain-language clinical narratives via the Claude API. Parents get a pediatrician-style summary they can share at appointments — not raw data, but insight.
screenshot · mobile
04
Mobile-Native Tracking
Built with React Native for iOS and Android. Quick-log feeds with one tap. Timer for nursing sessions. Weight entry with unit conversion. Push notifications for feeding schedules. Designed for one-handed use at 3 AM.
screenshot · mobile
Technical Decisions
Build vs. configure.
1
React Native over Flutter or native Swift/Kotlin
Cross-platform from day one. The app needs to work on whatever phone the parent has. React Native with Expo gives hot-reload development, OTA updates, and a single codebase. Trade-off: some native animations are harder, but the monitoring use case is data-heavy, not animation-heavy.
2
Custom SVG over Chart.js or D3
Growth charts need WHO percentile overlays, custom annotations, and baby-specific baseline rendering. No charting library supports this out of the box without fighting the abstraction. Raw SVG in React gives pixel-level control and smaller bundle size.
3
Claude API for narrative generation
Parents don't want to interpret charts — they want to know "is my baby okay?" The Claude API takes structured feeding/weight/vitals data and returns a clinical-style summary. Prompt engineering ensures medical accuracy without diagnosis. The narrative is a communication tool, not a diagnostic one.
Stack
React Native
Expo
TypeScript
FastAPI
Python
Claude API
SQLite
Try it.
Open Nourish →
← Back to all projects