Overview
Dive is a recreational scuba-planning client project that combines trip-planning features with a machine-learning safety profile and recommendations.
Challenge
The app brings together diver inputs, planning data, safety classification, cloud data, and location services. Those capabilities needed to remain one coherent mobile workflow rather than separate technical demonstrations.
My role
I owned and implemented the project end to end. My work covered the Android application, Python safety-classification component, API integration, Firebase services, maps and location features, and the delivered planning experience.
What I built
I built the Android client with Java, Retrofit, and Gson; used Firebase for authentication, realtime data, and storage; integrated Google Maps, Location, and Places; and connected the Python classifier that supplies the safety-profile behavior.
Architecture
Android Java client → Retrofit API → Python / scikit-learn safety classifier
Android Java client → Firebase Authentication / Realtime Database / Storage
Android Java client → Google Maps / Location / Places
Results
The delivered client project combines dive planning, safety classification, recommendations, account and data services, and location-aware mobile features in one application.
Lessons
Owning the full project made the boundaries between model behavior, service contracts, and mobile experience visible. The useful result was the integrated planner, not any one library in isolation.