trAIning

Using Tensorflow's Teachable Machine, trAIning uses AI and machine-learning to create a full-experience workout application. It is able to track user activity, count workout reps, display interactive analytics, and host social interaction between users.

Tech stack

This application is written in Typescript. It uses Firebase for the back-end and FireStore for the database. React, Less, Ant Design, Dayjs, React Webcam, Craco, TensorFlow, and TeachableMachine on the front-end. trAIning was deployed on AWS.

Team members

• Natasha Vermeulen | LinkedIn | Github

• Jaime Diez Rodriguez | LinkedIn | Github

• Kristen Hickey | LinkedIn | Github

• Edu Minguez | LinkedIn | Github

Screenshots

Main Features

• Workout creation
• Rep and Set counting
• Save workouts
• Add friends and send challenges
• Interactive Analytics

Instructions

Set up the backend on Firebase

• Create a firebase project.
• Add Authentication and enable login with email and password.
• Create a database and choose a Cloud Firestore location.
• Start two collections: profiles, workoutsDb.

Start the Application

• Clone this repository.
• Navigate to ./client and run npm i.
• To find your custom values on Firestore, go to Project Settings on the Project Overview dropdown menu.
• Create a .env in ./client with the following structure and add your custom values:

REACT_APP_APIKEY=
REACT_APP_AUTHDOMAIN=
REACT_APP_PROJECTID=
REACT_APP_STORAGEBUCKET=
REACT_APP_SENDERID=
REACT_APP_APPID=

• In ./client start the website with npm start

Demo video:

Video