NutritionChecker

Inspiration

Over the past few months, the world has been greatly affected by the coronavirus. Hundreds of millions of people have been forced to keep at home and shelter themselves from the outside world and although doing so has kept many individuals safe from the coronavirus, an increasing number of people are leading sedimentary lifestyles. Few of our team members had this common problem which is: they are not exercising or living an active lifestyle, but are still eating the unhealthy food that they could've eaten before quarantine when they were exercising and moving around more. With many individuals looking to stay at home and protect themselves from the coronavirus, which increases the likelihood of living a sedimentary lifestyle, the only way to stay fit and healthy is for them to control the food they eat. The saying, "We are what we eat", is the mindset our team carried throughout this event as we think that the saying accurately describes the importance of eating healthy food especially during a time like this. Just to sum it up, we chose to make the app we made because we understand how important the food we consume is, and the extent to which it can help or harm people; making healthier food choices is incredibly beneficial and is something that people should try to do.

What it does

The app is meant for shoppers who want to make healthy choices while buying food for themselves and their family members. The app calculates a healthiness index for each food item the user takes a picture of(the user takes a picture of the nutrition label of each item) based on data taken from numerous reputable health-related sources. Users can also store various shopping trips with an average healthiness score that is calculated by taking the average score of all the items part of that shopping trip. This enables the user to go through their shopping trip history and understand what they've been buying to hopefully make healthier choices or continue making healthy choices while shopping in the future.

How I built it

We used Java in Android Studio to build a fully functioning android app. For the OCR(used to get ingredients from the nutrition label), we used the Android ml kit.

Challenges I ran into

There were a handful of challenges that we ran into. First and foremost, all four of us on the team did not have too much Android development experience coming into this hackathon, so we had to learn a lot during this event to create the app. Apart from the lack of Android development experience, we had to ensure that the quality of the image that we were sending to the OCR model was high to be able to get the ingredients from the picture that the user takes. That process took some time because we had to understand how to send full resolution images between different activities without getting the dreaded "Transaction too large" error that we had been getting for some time.

Data persistence was also an issue that took some time to understand. We had to store many different ArrayLists and HashMaps that we wanted to stay constant throughout the app's lifetime and because of our lack of experience, it took some time to understand how to write variables to non-volatile storage.

Accomplishments that I'm proud of

We are all extremely proud of creating a fully functioning app in the limited time that we had. Although we all knew Java, we were a bit new to Android Studio, and being able to learn and actually create an app is something we're all proud of. In addition, being able to overcome the obstacles that we faced such as creating persistent data or working with high-resolution images during the development process is an incredibly satisfying feeling.

What I learned

We learned about various mobile app development and mobile design techniques while also learning the intricacies of Android Studio and android development as we worked more and more on the project. We also learned about the fundamentals of machine learning while working on this project in order to implement and improve our OCR quality.

What's next for NutritionChecker

We were hoping to have time to implement a feature in which the user can select a number of diseases or allergies that they have, and we can alert them if the food that they are buying is particularly detrimental to their health. We will continue to work on this app to see how much better we can make it, and we will definitely be adding this feature.