Ever had difficulty finding a roommate? Tired of endless roommate search across WhatsApp, Telegram, Facebook? Just click and find your perfect UMass roommate match at UMate!
UMate is a web application designed to simplify the process of finding compatible roommates. With UMate, users can easily create a profile, set their preferences, and find potential roommates using our interactive interface similar to dating apps. By using a recommendation model, UMate suggests the top matches based on compatibility factors like the duration of stay, lifestyle, budget, dietary preferences, and more.
- Features
- Tech Stack
- Installation
- Usage
- API Endpoints
- Contributing
- License
-
User Authentication: Secure login and sign-up with:
- Password hashing
- Email verification
- Unique username generation
-
User Profiles: Customizable profiles with:
- Budget
- Ethnicity
- Age
- Preferences ( dietary, location-wise, gender, identity, course, age, etc.)
- Smoking and drinking habits
- Lifestyle choices (quiet hours, guests, etc.)
- Duration of stay
- Personal bio
-
Recommendation System: UMate's recommendation system utilizes cosine similarity to measure the similarity between user profiles. This approach ensures accurate and personalized matchmaking with:
- Cosine similarity-based matching
- User profiles represented as vectors
- Normalized attribute values
- Algorithm Steps:
- Data collection
- Preprocessing
- Similarity calculation
- Matrix construction
- Top-K matching
- Recommendation generation
- Future Enhancements
- Additional attributes
- Weighted cosine similarity
- Collaborative filtering
-
Filtering Options: Users can filter potential roommates post recommendations by budget, dietary restrictions, lifestyle preferences, and more.
-
Friendly Interface: Inspired by dating apps, users can match through potential matches to find their ideal roommate.
- Frontend: React
- Backend: Flask with Python
- Database: MongoDB
- Cloud Storage: AWS S3 Bucket (Image Storage)
To get the project up and running locally, follow these steps:
Prerequisites
- Node.js and npm for the React frontend
- Python 3 for the backend
Clone the Repository
git clone https://github.com/your-username/UMate.git
cd UMateFinder
Setting Up the Backend (Flask API) Navigate to the backend directory:
cd backend
Install the Python dependencies:
pip install -r requirements.txt
Start the Flask server(By default runs the app on port 5000):
flask run
Setting Up the Frontend (React) Navigate to the frontend directory:
cd frontend
Install the npm dependencies:
npm install
Start the React development server:
npm start
The frontend should now be running on http://localhost:3000 and the backend on http://localhost:5000 on Windows. However, on MacOs, port 5000 is occupied with some other process. So run the backend on MacOs on a port other than 5000. Use the following command to run the backend on port 5050:
flask run --port 5050
Open the frontend application in a browser at http://localhost:3000. Sign up for an account or log in if you already have one. Set up your profile by specifying preferences such as budget, dietary restrictions, lifestyle choices, and desired duration of stay. Use the interface to browse recommended roommate matches. Use the filtering options to refine your search based on your criteria. Once you find a match, connect with potential roommates and discuss moving plans!
Here’s a summary of the main API endpoints:
User Authentication
- POST /user/insert: Create a new user account
- POST /user/validate: Log in to an existing account
Profile Management
- GET /user-details/: Fetch user preferences
- POST /user-details/: Add user preferences
Roommate Matching
- GET /rs/top-match: Fetch the top 5 recommended roommates based on user email
- GET /rs/other-mates: Fetch all users except the top recommended roommates for a user
Other APIs
- POST /user/favourites: Mark or unmark a roommate favourite for a user
- GET /user/favourite-roommates: List all favourite roommates and their details for a user
- GET /all_users/data: Return all users data except for password Note: Ensure that the backend server is running on http://localhost:5000 (for MacOS - try a different port other than 5000) and the frontend server on http://localhost:3000 for a smooth connection.
-
CORS Issues: We encountered Cross-Origin Resource Sharing (CORS) problems while integrating the frontend with the backend. Ensuring secure communication between the React frontend and Flask backend required configuring CORS policies correctly, especially for handling API requests smoothly.
-
Flask Integration for the Recommendation Model: Integrating the recommendation model into the Flask backend was challenging. This involved setting up efficient API endpoints that could handle and process recommendation logic, ensuring real-time data handling and accurate matching.
-
Data Processing for Recommendations: Preparing the data for recommendation was complex, requiring thorough preprocessing of user profiles and preferences. It was essential to clean, standardize, and transform data to accurately align with the compatibility factors used by the recommendation model.
-
Deploy it on AWS EC2
-
Enhanced Matching Algorithm: We plan to refine the recommendation model by incorporating more personalized factors, such as shared hobbies or specific living preferences, to improve match accuracy and user satisfaction.
-
Social Media Integration: Just like Bumble, integrating social media handles into the user profiles could allow users to verify their personalities and lifestyles, providing a more holistic view of potential roommates.
-
Real-Time Chat Feature: Adding a real-time chat feature would enable users to communicate with potential roommates directly within the app, streamlining the process of finalizing a match and moving forward with plans.
-
Roommate Reviews and Ratings: To build more trust in the roommate search process, we could implement a review and rating system where users can leave feedback on their previous roommates, helping future users make informed decisions.
Contributions are welcome! If you'd like to improve UMate, please fork the repository, make your changes, and submit a pull request. Ensure your changes align with the project's goals and are thoroughly tested.