Building PlanetKu: Android UI, Maps, and ML Integration
My Android development contribution to a six-member Bangkit capstone, from account flows to maps and the ML team's classifier.
The product and my role
Waste management is a significant challenge in many urban areas. The goal of Planetku was to create a mobile application that empowers users to manage their waste more effectively and sustainably. The app aims to solve problems like incorrect waste sorting, lack of awareness about recycling value, and difficulty in finding nearby waste banks.
I worked on UI/UX, login and registration, Google Maps, backend integration, and bringing the ML team's model into the Android application. Model development and evaluation belonged to the ML team.
Connecting the application
Connecting UI screens, account access, map locations, backend responses, and the ML team's classification output into one coherent Android flow.
- Waste classification through ML integration
- Google Maps waste-bank locations
- Login and registration
- UI/UX and backend integration
Working across disciplines
Integration required coordinating data formats and interface states with the backend and ML teams.
Results and attribution
The ML team reported 90% classification accuracy. My contribution was Android development and model integration; a reproducible model evaluation is not included here.
Integrating a team's ML output involves more than displaying predictions: the Android flow must coordinate account state, backend data, maps, and clear feedback for users.