Travelmate helps people connect with their cultural identities and build real-life communities, no matter where they are. It’s a platform for finding and hosting events, sharing traditions, and making meaningful connections with others who have similar backgrounds and interests.
Being away from home can feel isolating. Travelmate was built to bridge that gap by connecting people with shared cultural backgrounds. Through self-hosted events and a smart recommendation system, our platform helps users build genuine communities. Whether it’s sharing a meal, celebrating a tradition, or just having a conversation, Travelmate makes it easier to find a piece of home, wherever you are.
- Event Hosting and Discovery: Create and find cultural events hosted by community members.
- Personalized Recommendations: A recommendation engine suggests people and events based on location, interests, cultural background, and more.
- Direct Messaging: Connect and chat with new friends and acquaintances.
- User Profiles: Create a profile that showcases your background, interests, and what you're looking for in a community.
Travelmate is powered by a robust backend that combines a custom-built PostgreSQL database with a sophisticated recommendation engine.
-
PostgreSQL Backend: Our database is the heart of the platform, handling everything from user authentication and direct messaging to event management. We designed a custom schema to ensure data integrity and efficient querying, allowing us to manage complex relationships between users, posts, and conversations.
-
Recommendation Engine: The recommendation system is designed to provide relevant and diverse suggestions for both people and events.
- People Recommendations: We generate a candidate pool from friends-of-friends and users with similar attributes, then calculate a compatibility score to rank the suggestions.
- Post/Event Recommendations: We use a hybrid approach. First, a deterministic scoring algorithm ranks posts based on a variety of factors. Second, we use vector embeddings (via
pgvectorand Hugging Face) to represent users and posts, and rank by cosine similarity. The final recommendations are a weighted combination of both methods.
This creates a powerful loop: the data from user interactions and profiles is fed into our recommendation algorithms, and the resulting recommendations are stored back in the database to be served to the users, creating a constantly improving and personalized experience.
├── backend/ # FastAPI backend with recommendations API
├── frontend/ # Vite React frontend
└── scripts/ # Development scripts
To get a local copy up and running, follow these simple steps.
- Node.js and npm
- Python 3.8+ and pip
- PostgreSQL
-
Clone the repo
git clone https://github.com/exekis/hacks13.git cd hacks13 -
Database Setup
Make sure you have PostgreSQL installed and running. Then, run the setup script:
./scripts/db-setup.sh
This will create the database, tables, and seed it with sample data.
Note on Database Connection: The backend connects to PostgreSQL using environment variables. You can create a
.envfile in thebackenddirectory to configure your database connection. If you don't, it will use default values.# backend/.env DB_NAME=hacks13 DB_USER=your_username DB_HOST=localhost DB_PORT=5432 DB_PASSWORD=your_password -
Setup the Backend (Terminal 1)
Create and activate a virtual environment:
cd backend python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate pip install -r requirements.txt cd ..
Then, run the backend server using the script:
./scripts/backend.sh
-
Setup the Frontend (Terminal 2)
cd frontend npm install npm run dev
The frontend will be available at http://localhost:5173 and the backend at http://localhost:8000.
- A mobile app version of Travelmate.
- Exploring more advanced recommendation algorithms like collaborative filtering.