
Predicting blood-sugar levels of type 1 diabetics using machine learning
Keeping diabetics blood-sugar levels in the safe range can be a difficult and daunting task. Using modern machine learning techniques, Vampïr aims to make this task a little bit more manageable so you can focus more on doing the things you love to do.
This project was constructed with a group of friends at a local 24-hour hackathon (Mountainland Code Camp). Vampïr won 1st Place in the business category.
A little bit later, our work on Vampïr won my friends and I the 2018 High School Innovator of the Year Award. The iOS and Android apps are both open source and available on GitHub (see below).
Features
- Predicts blood glucose levels 20 minutes away
- Powered by machine learning
- Integrated with your Dexcom account
- Get notified when your blood sugar levels are predicted to go low or high.
Skills Used
- Kotlin
- Material Design UI
- Machine Learning
- Android Development
- RESTful APIs
- MySQL
- Android Architecture Components
- MVVM App Architecture
- Business Management
- Time Management
- Presentation
Get Tempo