Diabetes Viz
Project Infobox 
| |
|---|---|
| Self researcher(s) | Peter Kok |
| Related tools | blood glucose monitor, Google Fit |
| Related topics | Chronic disease, Blood glucose tracking, Insulin intake, Activity tracking |
Builds on project(s) |
|
| Has inspired | Projects (0) |
| Show and Tell Talk Infobox | |
|---|---|
| Featured image | 
|
| Date | 2017/06/18 |
| Event name | 2017 QS Global Conference |
| Slides | Diabetes-viz.pdf |
|
This content was automatically imported. See here how to improve it if any information is missing or out outdated. |
Diabetes Viz is a Show & Tell talk by Peter Kok that has been imported from the Quantified Self Show & Tell library.The talk was given on 2017/06/18 and is about Chronic disease, Blood glucose tracking, Insulin intake, and Activity tracking.
Description[edit | edit source]
A description of this project as introduced by Quantified Self follows:
Peter Kok is a diabetic with a background in data visualization. In this video, he discusses how he has gathering data from his insulin pump, continuous glucose sensor and activity tracker. He also developed a visualization tool to make sense of these diverse data sources.
Video and transcript[edit | edit source]
A transcript of this talk is below:
Peter Kok - Diabetes Viz
I’d like to share about what it means to have diabetes type I, what data you can track, what visualizations you can use to make sense of the data and how I use that data to manage my diabetes. So diabetes type I is a condition where you have, where you don’t produce insulin and you need insulin for glucose to enter your cells where it can be burnt and used as energy. If you don’t have insulin, the sugar accumulates and can lead to all kinds of complications. You can manage diabetes by using insulin, administrating insulin through injections or through an insulin pump and you determine the dosage based on your glucose level that you measure yourself. And the target is to keep the glucose levels in a tight range, and anywhere outside that range there’s a risk of long-term complications or even worse. So there’s a lot of factors affecting glucose levels and the list is a bit longer than what you see here, so that makes it a very complex disease. The good thing about it is there’s a lot of data to track. So what I track, I don’t track all the variables now, so I’ve a three style lever like the glucose sensor that Yerin also showed yesterday, you know which measures my glucose levels under the skin. I can read it with a reader and I export the data to a CSV file. Then I use an insulin pump to administer the insulin, but this also keeps track of when you administer insulin and how much. And I can also export that data to a CSV file. As an activity tracker at the moment I use Google Fit and I also export that to a CSV file. So the only important thing I don’t track now is food intake and that’s still on my to-do list. With this data, being a software engineer, I created this custom web application that brings together these data sources to make sense of this data. It’s written in JavaScript. It runs on your bar so you can try it yourself, it’s still a little bit buggy. But it gives a nice impression of what’s possible. I’m using a library called DC.js. it stands for dimensional charting and it’s very useful for creating multidimensional visualizations. So how did I get there? So when I just plot the raw data I get something like in the top left corner, so you cannot make sense of that obviously. So you need to aggregate it, so in this case I took the average for each day, but then you lose some information and that’s because of your variability of your data is also an important factor in how you’re doing. So you can’t see this very well, but I added this band showing the variance. So if you zoom in, it automatically switches to a different level of detail. So instead of the average you see the raw values. So you can get back to your original resolution of the data. That is all very nice, but you also want to have some context so I added this graph at the bottom that shows actually the context of where you are in your timeline. So over a period of about two months, here you see about five days and you can also use this to move around your data to make selections and stuff like that. So DC.js is also great for creating new dimensions, so you can aggregate the data by hour of day or by day of week. And you can also make selections in this and it automatically updates the other charts with this selection. So here you see I have a curious spike on Thursday evening and that coincidently corresponds to the time we had drinks after work, so that’s interesting to see that. So I also added activity and insulin. So activity is in the purple bars and insulin is in the kind of orange bars at the top. I’m still not fully satisfied with this kind of visualization, but at least it shows you some coincidence of activity and the drop in sugar levels. What I find myself doing still is going back to my calendar to see what I was actually doing during that time. So that’s something I might also want to include in this visualization. So there’s still a lot on my wish list, so I would like to include more variables like food intake. I’d like to automatically detect better and find out what are the best patterns, and automate the synchronization of my data. So the affects this has on my diabetes is that I’m now able to identify these patterns and look more closely at what is happening there and how can I improve my lifestyle. So when I go for a run should I use less insulin or at what times, but yes stuff like that. But also it really helps me to keep motivated. I like to keep my glucose levels within this tight range. So I like to say seeing is believing so it really motivates me to act on the data. So that’s what I wanted to share.
Thank you.
About the presenter[edit | edit source]
Peter Kok gave this talk.