NutriCorr

Our expected UI
The terminal output of prediction score after data processing

Inspiration

Research shows that maximum people face mental or physical health problems due to their unhealthy daily diet or ignored symptoms at the early stages. This app will help you track your diet and your symptoms daily and provide recommendations to provide you with an overall healthy diet. We were inspired by MyFitnessPal's ability to access the nutrition information from foods at home, restaurants, and the grocery store. Diet is extremely important to the body's wellness, but something that is hard for any one person to narrow down is: What foods should I eat to feel better? It is a simple question, but actually very hard to answer. We eat so many different things in a day, how do you know what is making positive impacts on your health, and what is not?

What it does

Right now, the app is in a pre-alpha phase. It takes some things as input, carbs, fats, protein, vitamins, and electrolyte intake in a day. It sends this data to a Mage API, and Mage predicts how well they will feel in that day. The Mage AI is based off of sample data that is not real-world data, but as the app gets users it will get more accurate. Based off of our data set that we gather and the model type, the AI maintains 96.4% accuracy at predicting the wellness of a user on a given day. This is based off of 10000 users over 1 day, or 1 user over 10000 days, or somewhere in between. The idea is that the AI will be constantly learning as the app gains users and individual users enter more data.

How we built it

We built it in Swift using the Mage.ai for data processing and API

Challenges we ran into

Outputting the result on the App after the API returns the final prediction. We have the prediction score displayed in the terminal, but we could not display it on the app initially. We were able to do that after a lot of struggle. All of us made an app and implemented an API for the very first time.

Accomplishments that we're proud of

-- Successfully implementing the API with our app -- Building an App for the very first time -- Creating a model for AI data processing with a 96% accuracy

What we learned

-- How to implement an API and it's working -- How to build an IOS app -- Using AI in our application without actually knowing AI in depth

What's next for NutriCorr

--Adding different categories of symptoms -- giving the user recommendations on how to change their diet -- Add food object to the app so that the user can enter specific food instead of the nutrient details -- Connect our results to mental health wellness and recommendations. Research shows that people who generally have more sugar intake in their diet generally stay more depressed.