Get personalized health predictions and recommendations based on your daily activities.
The idea didn't come from a lab or a lecture. It came from a late-night brainstorming session, where ideas bounce around like lofi beats on loop. We started talking about the Internet of Things (IoT)—the network of connected devices that's quietly powering our homes, cities, and now, healthcare.
Sleep is tied to everything—stress, chronic conditions, mental health. If you improve sleep, you improve health. And here's the kicker: Wearables already track sleep-related data like heart rate, motion, and sleep stages. The data is there; it just isn't being used to its full potential.
We used a dataset by Olivia Walch and her team. It had everything we needed—heart rate, motion, and gold-standard sleep labels. It was clean, reliable, and a perfect starting point for our model.
For the modeling, we turned to PyCaret, a low-code library that let us quickly compare machine learning algorithms. KNN stood out for its simplicity and speed, which are key for real-time applications.
We're working on connecting this to APIs from Fitbit, Apple Watch, and WHOOP. Imagine real-time data flowing into a system that predicts how tonight's sleep could change based on your daily activity.
The future of healthcare is predictive, personalized, and participatory. With wearable-based digital twins, we could:
Simulate real-life scenarios to accelerate research and drug trials
Enable clinicians to predict and prevent complications with real-time monitoring
Give individuals the tools to take control of their health in ways that are tailored and actionable