Mobile development
Cowspace
Coworking Space Recommendation App
About the project
A mobile application that helps students discover suitable places to study or work based on their preferences, budget, and required facilities.
Cowspace addresses the difficulty of finding and comparing suitable study or workspaces across multiple platforms. Developed using Flutter, FastAPI, and MySQL, the application integrates a hybrid recommendation system to generate personalized recommendations from 35 curated locations in South Jakarta.

Inside the experience
Personalized Top-5 recommendations
Preferences based on place type, budget, and facilities
Detailed place information, ratings, and locations
User ratings and recommendation history
Behind the system
Full-stack mobile architecture
Integrated a Flutter mobile interface with a FastAPI backend and MySQL database through RESTful APIs, separating the user interface, recommendation logic, and database operations.
Hybrid recommendation engine
Combined Content-Based Filtering (20%) and User-Based Collaborative Filtering (80%) to generate ranked recommendations using user preferences and rating patterns. The content-based component also supports recommendations for new users without rating history.
Project outcomes
93.12% Precision@5
Achieved by the hybrid recommendation model during offline evaluation using 5-fold cross-validation.
27/27 Test Scenarios Passed
Successfully completed 5 unit tests, 4 integration tests, and 18 black-box test scenarios.
94.82% User Acceptance Score
Achieved through usability and functionality evaluation involving five participants.
Interface gallery
Select an image to explore the details.