Predicting the probability of nocturnal hypoglycemic events with probabilistic forecasting.
13
stars
268
commits
Jupyter Notebook
primary language
Sep 5, 2026
updated
README
Probabilistic Forecasting of Nocturnal Hypoglycemia
Find our documentation here: Documentation Find a working summmary of our results here: https://blood-glucose-control.github.io/nocturnal-hypo-gly-prob-forecast/
About the Project
Fear of nocturnal hypoglycemic events continues to be one of the most significant contributors to diabetes distress. Developing improved nocturnal hypoglycemic forecasting techniques would likely improve diabetes distress (D. Ehrmann et al., 2024), but even short 2-hour blood glucose level (BGL) forecasting remains challenging (H. Nemat et al., 2024). Forecasting BGL over an 8-hour time window does not seem feasible with SOTA techniques, likely due to the limited scale of most BGL time series datasets. The most popular open-source T1D dataset, OhioT1DM, only contains 12 patients over eight weeks (C. Marling,2020). Our student volunteer-based WAT.ai Blood Glucose Control Design Team, in partnership with Gluroo, aims to reframe the forecasting problem into a big data probabilistic forecasting problem.
Probabilistic forecasting is about producing low and high scenarios, quantifying their uncertainty, and delivering expected ranges of variation (T. Gneiting et al., 2014, F. Kiraly et al., 2022). Our view is that this framing is more technically feasible yet will be just as valuable in alleviating diabetes distress related to fears of nocturnal hypoglycemic events.
Methods
Our research will consider months of data from thousands of de-identified Gluroo patients. With skTime (F. Kiraly, 2024) we will evaluate various interval, quantile, variational, and distributional forecasting techniques to maximize the sharpness of our predictive distributions (T. Gneiting et al., 2014,, F. Kiraly et al., 2022). We will contrast our forecasts over various categories, including age groups, gender, years with diabetes, and weekend vs. weekday forecasting.
Project Setup
For instructions on getting started with our project and the commmit/PR procedure please see our wiki article: Getting Started Developing with this Repo
License
This project is licensed under a custom research license based on CC BY-NC 4.0. The software is freely available for academic institutions and non-profit research organizations, but commercial use requires a separate commercial license.
See LICENSE for full details.
For commercial licensing inquiries, please contact: christopher@gluroo.com
License Summary
- ✅ Free for research - Academic institutions and non-profit research organizations
- ✅ Attribution required - Please cite Blood-Glucose-Control in your work
- ❌ Commercial use restricted - Requires separate commercial license
- 📧 Commercial licensing - Contact us for business use
Contributors
Similar repos
Languages
Jupyter Notebook
90.4%
Python
9.0%
Predicting the probability of nocturnal hypoglycemic events with probabilistic forecasting.
13
stars
268
commits
Jupyter Notebook
primary language
Sep 5, 2026
updated
README
Probabilistic Forecasting of Nocturnal Hypoglycemia
Find our documentation here: Documentation Find a working summmary of our results here: https://blood-glucose-control.github.io/nocturnal-hypo-gly-prob-forecast/
About the Project
Fear of nocturnal hypoglycemic events continues to be one of the most significant contributors to diabetes distress. Developing improved nocturnal hypoglycemic forecasting techniques would likely improve diabetes distress (D. Ehrmann et al., 2024), but even short 2-hour blood glucose level (BGL) forecasting remains challenging (H. Nemat et al., 2024). Forecasting BGL over an 8-hour time window does not seem feasible with SOTA techniques, likely due to the limited scale of most BGL time series datasets. The most popular open-source T1D dataset, OhioT1DM, only contains 12 patients over eight weeks (C. Marling,2020). Our student volunteer-based WAT.ai Blood Glucose Control Design Team, in partnership with Gluroo, aims to reframe the forecasting problem into a big data probabilistic forecasting problem.
Probabilistic forecasting is about producing low and high scenarios, quantifying their uncertainty, and delivering expected ranges of variation (T. Gneiting et al., 2014, F. Kiraly et al., 2022). Our view is that this framing is more technically feasible yet will be just as valuable in alleviating diabetes distress related to fears of nocturnal hypoglycemic events.
Methods
Our research will consider months of data from thousands of de-identified Gluroo patients. With skTime (F. Kiraly, 2024) we will evaluate various interval, quantile, variational, and distributional forecasting techniques to maximize the sharpness of our predictive distributions (T. Gneiting et al., 2014,, F. Kiraly et al., 2022). We will contrast our forecasts over various categories, including age groups, gender, years with diabetes, and weekend vs. weekday forecasting.
Project Setup
For instructions on getting started with our project and the commmit/PR procedure please see our wiki article: Getting Started Developing with this Repo
License
This project is licensed under a custom research license based on CC BY-NC 4.0. The software is freely available for academic institutions and non-profit research organizations, but commercial use requires a separate commercial license.
See LICENSE for full details.
For commercial licensing inquiries, please contact: christopher@gluroo.com
License Summary
- ✅ Free for research - Academic institutions and non-profit research organizations
- ✅ Attribution required - Please cite Blood-Glucose-Control in your work
- ❌ Commercial use restricted - Requires separate commercial license
- 📧 Commercial licensing - Contact us for business use
Similar repos
Contributors
Languages
Jupyter Notebook
90.4%
Python
9.0%