Description as a Tweet:

PostuReMinder - Improve Your Overall Health One Posture Check at a Time!

Inspiration:

With the COVID-19 pandemic, many of us are stuck at home and as a result, we spend hours staring at a screen which is extremely unhealthy. This inspired us to create something that can help keep people healthy by reminding them to drink water, maintain good posture, and stretch when working on a computer.

What it does:

Our project uses machine learning facial recognition technology to detect whether a person is slouching on their chair, and sends them notifications to fix their posture when bad posture is detected. It also includes an analytics page, which helps show the user their improvement or deterioration in posture over time. Additionally, we also have Water and Stretch Reminder buttons, which the user can customize to determine how often they want to be notified to drink water and stretch.

How we built it:

Our plan of attack for this hack was to break it apart into reasonable chunks and split those evenly. We always made sure the others were aware of what we were each currently working on. Being on a call the whole time was extremely helpful as we were able to provide immediate assistance anytime someone was facing a blocker.

Technologies we used:

• HTML/CSS
• Javascript
• AI/Machine Learning

Challenges we ran into:

One of our biggest challenges was just learning HTML, CSS, and Javascript.

For the machine learning portion, we wanted to show to the user what was going on in the backend, but this required running a ML model in real time--and we wanted to do it in the browser. So we had to work with async Javascript code and find suitable parameters which balanced performance and power consumption.

Accomplishments we're proud of:

First off, we are very excited to build this website that anyone can use. But beyond that were some of the little things - understanding how JSON is formatted, how asynchronous functions work, and how HTML, CSS, Javascript interact with each other.

What we've learned:

Coming into this hackathon, our team didn’t have much knowledge in web development. But by struggling together, we were able to explore the ins and outs of creating webapps. By the end, we can safely say we can build some solid full functioning websites. We initially thought our idea might be a bit too difficult to implement, but we came to realize that once we put our minds to it, anything is possible.

What's next:

Although we accomplished all of what we had planned for this hackathon, there are still many features to add on or to improve to. To name a few, we plan on improving the machine learning model to make it more efficient, adding more statistics visuals, and overall adding more customization for each of the three reminders.

Built with:

We worked together in real-time using Repl.it and Github. We used a vanilla HTML, Javascript, CSS front-end and a Javascript machine learning back-end (face-api.js). We used assets from Current and other libraries. The Website was registered with Domains.com and hosted using Netlify (https://imposture.tech/).

Prizes we're going for:

• Best Web Hack
• Funniest Hack
• Best Domain Name
• Best Machine Learning Hack
• Best Healthcare Hack
• Best Beginner Software Hack
• Best Beginner Web Hack

Jeffrey Liu
Steven Zhang
Alan Wu
David Huang

Table TBD