💡Inspiration

Because of the pandemic, people realize the importance of caring for their health more and more. There is now an ongoing demand for ways to train more effectively and conveniently since people cannot go out and need to do social distancing. Further, even after the pandemic, the thought of training conveniently anytime or anywhere is now imprinted in many customers’ minds.

These facts all point to the need for creating an AI-powered personal fitness trainer site that can assist customers anytime or anywhere they want. The AI fitness trainer can understand the customers’ needs and is able to develop a well-planned training plan. The AI fitness trainer checks and improves customers’ fitness routines while making remote support accessible.

Because of the above requirements, we designed our site, Fitnessify, for customers to check and improve their fitness routine. The solutions our team came up with were to build an AI powdered personal fitness trainer site with an AI chatbot that advises customers about training tips and metrics about the training progress of the customers. We also have automatic replies and phone calls through text conversations with the customers. We also provide the users with their weekly progress reports and send them notifications reminding them about their routine.

💻What it does

An AI powdered personal fitness trainer site with an AI chatbot that advises customers about training tips. Customers can see personalized fitness routine, diet plans, and their weekly progress stats (metrics about training progress of the customers) on their logged-in dashboards. Customers also have an AI Fitness assistant which assists them in their workouts. There is a posture analyzer and a Yoga posture detector built on AI just to assist the customers. We also have automatic communication with customers through Twilio. Customers can send a text message and will get automatic text messages/phone calls for more assistant.

🔨How we built it

  • Python
  • Azure bot
  • Microsoft Azure
  • OpenCV
  • Tensorflow
  • HTML
  • CSS
  • Streamlit
  • Figma
  • Tableau
  • Twilio
  • Godaddy
  • Github

🧠Challenges we ran into

We only had 2 days to prepare everything and being from different time zones was a bit problematic, yet everyone cooperated and we overcame this issue. Had some initial trouble with our Open-CV model as it was not showing proper results, then trained it for better. Collecting data for Yoga Posture Detector and Posture Analyzer took some time. Posture Analyzer dataset was created by us from scratch as it was not available on google. Faced an initial setback with the front end when an additional teammate came to the rescue and we tried our best to implement our idea.

🏅Accomplishments that we're proud of

  • We were able to create a website almost close to our Figma Prototype and set it up working.
  • We could successfully develop the AI models and deploy them to analyze poses and activities of customers accurately.
  • The chatbot functions smoothly.
  • We could set up the metrics and set up automatic replies and phone calls through Twilio.

🤓What we learned

Work more efficiently and handle different parts of the project. Keeping the deadline in mind and forming a proper pathway to finish the project on time. Working with various ML models and deploying it properly. Integrating various parts of the project which none of us were very experienced with.

🚀What's next for Fitnessify

We want to give chatbot the ability to suggest activities that are good for customer’s fitness training plan based on data analytics of historical data. We have plans of adding more AI models based on posture detection. Update Twilio to send a confirmed text message when a customer purchases a plan. We also want to donate 25% of the revenue collected by our programs in charity.

Built With

Share this project:

Updates