WellShift

• 

  Depicts the data inputs and the sentiment timeline chart.

• 

  Show inputed data.

Inspiration

This project is inspired by a personal struggle with mental health, designed as a way to track and monitor emotional well-being.

What it does

WellShift is an AI-powered full-stack mental health tracking application designed to help users monitor their emotional well-being through intelligent sentiment analysis. It transforms daily reflections into a visual timeline of emotional trends.

How we built it

WellShift was built as a modern, high-performance, full-stack application using: Frontend: ReactJS and TypeScript powered by ViteJS. Backend: Python with the FastAPI framework for high-speed asynchronous processing. AI/ML: A pre-trained RoBERTa Transformer model for nuanced sentiment analysis. Database: MongoDB Atlas for scalable, cloud-based data persistence. Deployment: GitHub Pages (Frontend) and Hugging Face Spaces with Docker (Backend).

Challenges we ran into

I faced challenges in optimizing the integration of the large RoBERTa Transformer model within the FastAPI server to ensure low-latency inference. Additionally, migrating from a local SQLite setup to MongoDB Atlas required a shift in data handling, and configuring CORS and environment variables across GitHub Pages and Hugging Face was a significant hurdle.

Accomplishments that we're proud of

Integrating a sophisticated RoBERTa Transformer model rather than relying on simpler, keyword-based NLP methods.

Creating a seamless full-stack pipeline from client-side input to a final, analyzed, and persisted data point in a cloud database.

Successfully containerizing the backend using Docker to ensure a consistent environment across local and cloud deployments.

What we learned

I gained significant hands-on experience in productionizing ML models within a web server environment (FastAPI). I also learned how to manage cross-origin resource sharing (CORS) in a decoupled architecture and how to securely handle sensitive credentials using environment variables and repository secrets.

What's next for WellShift

User Accounts: Implement secure authentication (JWT/OAuth) so users can save and sync their history across devices.

Enhanced Visualization: Implement a dedicated dashboard view for long-term emotional trends (monthly and yearly).

Personalized Feedback Loop: Introduce a feature that allows users to rate supportive nudges, enabling the system to refine messaging over time.

Mobile-First Deployment: Optimize the UI/UX for deployment as a Progressive Web App (PWA) to enhance accessibility.

Built With

• docker
• fastapi
• github
• huggingface
• mongodb
• python
• react
• roberta
• transformer
• typescript