0) Tech choices (beginner-friendly)
React Native + Expo → fastest to ship; runs on your phone with Expo Go.
Navigation: @react-navigation/native
Camera: expo-camera
GPS: expo-location
Maps (online now, switchable to offline): react-native-maps with OpenStreetMap
tiles
Image resize/preprocess: expo-image-manipulator
SOS (SMS composer): expo-sms
Local storage: @react-native-async-storage/async-storage
When you’re ready for offline ML
Use one of these (requires “Dev Client” or Bare workflow):
ONNX Runtime for React Native: onnxruntime-react-native (recommended)
or TensorFlow Lite: react-native-tflite / native module
You can start with a mock classifier (works today) and swap to ONNX/TFLite later without
changing the UI.
1) 15-Day Plan (3–4 hrs/day)
Day 1–2 – Install tools, create Expo app, set up navigation & screens
Day 3 – Camera preview + capture, basic UI polish
Day 4 – Image preprocessing; integrate mock classifier (returns top label)
Day 5 – GPS location + map with trek polyline (from local GeoJSON)
Day 6 – SOS screen: save contacts, compose SMS with live GPS link
Day 7 – Trek info module (static JSON: difficulty, time, distance)
Day 8–9 – Prepare dataset (50–100 species pilot), labeling & curation
Day 10–11 – Train quick model (see Python below), export ONNX/TFLite
Day 12 – Integrate ONNX Runtime classifier (or keep mock if short on time)
Day 13 – Cache tiles locally / plan offline maps (optional this sprint)
Day 14 – UX tidy, error handling, permission flows, airplane-mode testing
Day 15 – Final QA, screenshots, APK build (Expo EAS or classic dev build)