Sentire
Emotion Tracking Web Application
A full-stack web application for log & share emotions, track patterns over time, and analyze how activities influence mood securely over time.
Built using Svelte, Node.js (Express), and MongoDB & Neo4j, deployed using Docker and Nginx.
Why
The system tracks:
What you do → How you feel
Over time, it identifies patterns like:
- “Music improves mood”
- “Studying is associated with stress”
- “Gym correlates with positive emotions”
This enables data-driven self-awareness and behavioral insights.
Features
- User authentication
- Session-based authentication
- Emotion logging with color tagging
- Activity-based emotion tracking
- Social sharing with friends
- Analytics - activity vs emotion patterns
- Retrieve emotion history per user
- Manage active sessions (view, revoke single session, revoke all)
- Cross-origin support with secure CORS configuration
- Scalable backend with database abstraction
Architecture Overview
Tech Stack
- Frontend: Svelte + JavaScript
- Backend: NodeJS + ExpressJS
- Database: MongoDB + Neo4j
- Deployment & Infrastructure: Docker + Nginx (reverse proxy)
Why Neo4j
In order to perform activity vs emotion frequency analysis, we need to perform multiple many to many relation queries. These queries are computationally expensive in MongoDB due to individual document retrieval. Neo4j models these relationships natively as a graph, enabling efficient traversal and faster relationship-based queries
- Activity vs emotion frequency analysis
- Mood pattern detection
- Behavioral insights
- Relationship traversal
For example:
- Gym → 80% positive emotions
- Music → 90% positive emotions
- Studying → mostly negative
Environment Variables
Create a .env file in the backend root:
FRONTEND_URL=http://localhost:5173
MONGO_URL=mongodb://localhost:27017/dbname
NEO4J_URL=neo4j://username:password@localhost:7687
Using Docker
docker compose up -d --build
Running Locally
Backend
npm install
node index.js
Backend runs on:
http://localhost:3000
Frontend
npm install
npm run dev
