Objective digital data is scarce yet needed in many domains to enable research that can transform the standard of healthcare. While data from consumer-grade wearables and smartphones is more accessible, there is critical need for similar data from clinical-grade devices used by patients with a diagnosed condition. The prevalence of wearable medical devices in the diabetes domain sets the stage for unique research and development within this field and beyond. However, the scarcity of open-source datasets presents a major barrier for data-centric researchers.

The Glucose-ML project is an evolving collection of publicly available continuous glucose datasets curated, standardized, and maintained to accelerate data-centric research and development. This project seeks to:

• Bridge the data gap that imposes barriers for novice and established data science, machine learning, and AI practitioners seeking to contribute to health-relevant research.
• Provide a central hub for exploring, visualizing, and comparing real-world CGM datasets to support AI/ML research for diabetes prevention and care.
• Facilitate dataset transparency and enable joint use of disparate high-quality datasets to develop robust, reproducible, and generalizable algorithms that support health equitably.

Our team also collected and maintains the DiaTrend dataset which includes intensive longitudinal data from wearable medical devices (27,561 days of continuous glucose monitor data and 8,220 days of insulin pump data) from 54 patients with type 1 diabetes. This dataset is useful for developing novel analytic solutions that can reduce the disease burden for people living with diabetes and increase knowledge on chronic condition management in outpatient settings.

Interested?

Explore, discover, and access public continuous glucose monitoring datasets from the Glucose-ML website here: https://www.glucose-ml-project.com/.

‍Request access to the DiaTrend dataset through this link: https://doi.org/10.7303/syn38187184.