November 13, 2019 | |

With blood glucose levels often oscillating between high and low several times a day, the daily life of a diabetic can be quite the roller coaster. Inspired by a colleague living with Type I Diabetes, we want to use the skillsets and analytical knowledge at TD to create a tool that helps anyone trying to actively manage blood glucose and insulin levels do so in an informed manner. Our proposed open-source solution to blood glucose management aims to provide relevant and timely alerts and insights, aiming to make diabetes a smoother ride for those who live with it.

How do we do it?

We collected data on blood glucose, insulin and food intake and in combining these different sources, were able to develop a model that predicted future levels of blood glucose 30 minutes ahead of time. This type of prediction tool could help diabetics make more informed decisions in managing their diabetes on a daily basis.

The result was pretty much consistent with the real values.

Although the inputs for our model thus far included carbohydrate intake and previous data, our aim is to also include physical activity, heart rate, fat content of food, and other relevant inputs to improve our model. We believe that this will make the model more realistic and more usable for more people.

In order to take this next step, we need more data.

If you are interested in becoming a data donor to help us improve this open-source solution to managing diabetes, please get in touch with us via