Covid Tracker

October 2020 - October 2020

    Tags

  • Android
  • Express
  • Hackathon
  • Java
  • TensorFlow

About

A covid contact tracing app made during the hackathon Hack TUES 6.

Every minute every device makes a Bluetooth request to every other device nearby. Via ML it predicts which mode of transportation every user uses - car, still, walking, bus or train. If a user marks themselves as COVID-19 infected, it sends notifications to every other user, who may have been in contact with them. Every notification contains a map, which displays the exact point where the user has been in contact as well as in what circumstance they may have been infected.

Personally, I worked on the machine learning algorithm which predicts the mode of transport. The app gathers information from the gyroscope of the device and sends it to a Flask API with the model which makes its prediction and sends it to the main express server.

Due to circumstances surrounding our group's delivery, we were penalized and placed 4th in the hackathon.