TCuS Medical AI Web-App

Don't want to go to the doctors just yet and don't want to be an internet doctor? Try TCuS Health, a web-app using AI and a plethora of datasets to check all your symptoms and receive a diagnosis!

Comment

What was our Inspiration for TCuS Health?

As we started living away from home, it has been difficult to schedule doctor check-ups and attend them without stressing over losing time from homework and classes. Searching symptoms from the internet gave us so many different diagnoses-- I don't think I have cancer from a headache! Our solution is TCuS Health, a web-app that accurately records your symptoms to give a realistic diagnosis.

What it does

TCuS Health first displays a list of questions like "do you have a cough?" or "are you experiencing shortness of breath?". Then, each question is followed by a yes or no option or a enter value based on the question type. The app then takes the data and runs it through the AI Model we created to output if the user had a possible cardiovascular issue, stroke, cold, COVID, and influenza (flu).

How we built it

First we went through hundreds of medical Kaggle datasets to see what matched our needs all around. Then, we created a model by using SK Learn and TensorFlow that takes in numerical and categorical data which is what all of our suggestions are based from. Simultaneously, we created the front end on ExpoGo to have our questions shown to the user to display the final result.

Challenges we ran into

Some include...

  1. Connecting the AI model to the front end to the UI
  2. We had trouble overcoming a host problem with it showing an unresolved error.
  3. Lots of undefined errors :(
  4. The data was very very messy to clean and we had to hot encode a lot of values! 😞

Accomplishments that we're proud of

In these 24 hours...

  1. We cleaned and created 3 models based on medical data.
  2. Created a mobile app.
  3. We designed a easy to use UI.
  4. IT ACTUALLY WORKS AFTER BLOOD, SWEAT, & TEARS.

What we learned

  1. We make a great team-- we be-lung together :)
  2. We learned how to utilize the SKlearn library in conjunction with Tensorflow.
  3. We learned how to use Flask API
  4. How to connect the backend to the AI Model using dictionaries and connecting with the front end by using a ngroq api from our local host.

What's next for TCuS Medical AI Web-App

Whats next? Right now, the app focuses on respiratory and cardiovascular issues and diseases, and we plan on adding more diseases and disorders from more body systems. We also would love to add mental health issues in our app, so stay tuned so yall can get more med-tastic diagnoses! :)

Built With

Share this project:

Updates