Healthy Habits

Login Page
Main App
GitHub Repo

Summary

Our AI-integrated website is a service that runs in a user’s background and sends a notification for the user to consider a break if their emotions reach over a certain threshold for an amount of time. We do this by utilizing OpenCV and Mediapipe to detect faces and capture image data which is then sent off to Hume’s streaming API every 60 frames for processing. Their AI then sends back emotional data which we then use to gauge a user’s “negative” emotion levels. Our service also offers a report generation system if a user would like to receive comprehensive reports on their emotional levels over the past day or week.

Stack

Python, Reflex, HumeAI, CockroachDB, OpenCV

Target Audience

Our goal is to help workers who spend long hours at their computers or even students who have to spend long hours on their computers doing schoolwork or other activities. As computer science students who spend long hours on computers ourselves, we understand how sometimes students can spend endless hours on their work with no break, increasingly growing more agitated or frustrated as things don’t go as planned.

Mission

As Computer Science students, we certainly understand the pressure and difficulties working within a stressful environment for long hours. Our primary objective is to help people maintain their mental health through similar situations by suggesting breaks when they clearly need them. By promoting healthy work habits we aim to improve our client's efficiency, enabling them to succeed in an environment of their choice. We hope our tool will help many prevent burnout from endless grinding on their work and also help maintain their long-term mental health.

Challenges

Using new tools such as Reflex served as quite a difficulty as we were lacking on a lot of documentation and had to consult the Reflex staff. Furthermore, due to unfortunate difficulties on Reflex's end with Docker images, our project is unable to be hosted in production and our video streaming is severely laggy.

What We Learned

Throughout our time on the project, we were able to experiment with technologies both familiar and new to us. We expanded our skills in computer vision and full-stack development, and discovered a new framework to establish both a frontend and backend in python with Reflex.

Privacy

We do not collect data on users for profitable purposes but only to assist in providing analysis for users to better understand their health.

Future Plans

We plan to expand support to mobile apps for portable usage of our service and also plan to implement more features such as optimized health letters for physicians.