JETT

An AI occupany tracker for students looking for a good place to study!

Comment

Inspiration

We took inspiration from the presentation given by Buck Woody in the use of Azure AI.

What it does

The first AI (CAM) was trained to detect occupied tables and identify availability. The second AI (MAP) was a preset, Computer Vision, and was used to detect the distance two relative points on the campus map. Our back-end then takes these values to the front and displays the data in 2 unique presets:

  1. A bar of low, medium and high occupancy with colors green, yellow, and red, respectively.
  2. A percentage of the total people identified by the AI are divided by the amount of seats in a given room. We also have 3 additional features:
  3. Student study room booking
  4. The location to the nearest available study room feature; according to the schedule of the user** (TIME SPECIFIC: The schedule of the user obtained via the database)
  5. Photo upload feature (Try it on a non-preset room!)

How we built it

We trained 2 AI's, Custom Vision AI and Computer Vision AI. Utilized python to connect our back-end to the data processed by the AI. We then used React, HTML, Tailwind CSS and Javascript to visualize the data in the frontend.

Challenges we ran into

Azure was a brand new concept to our team. Being first time hackathon participants we ran into challenges and learning curves in the implementation of the AI.

Accomplishments that we're proud of

It runs! The fact that all of us are new to hackathons we were able to produce a working application. We are extremely proud of JETT!

What we learned

We learned extensively about the ins and outs of how Azure AI connects to different coding languages and frameworks, along with how to effectively use Git and Github.

What's next for JETT

JETT is definitely in its early stages and plans for the future are vast. To name a few things we hope to push for in the future being:

  • A fully functional chrome extension with canvas integration
  • Implement another AI to better scan the schedules and allow for photo upload
  • Using user location to show the nearest study space
  • Train the AI further better detect objects
  • Not be limited to the 4 buildings given via presets (Additional campuses included)
Share this project:

Updates