I received a CGM (Continuous Glucose Monitor) about two weeks ago via the Levels start-up. The package arrived a few days after ordering, and the initial setup was relatively easy and painless. I was able to easily export the data from the Levels website.
So how do the results look? The easiest way I can think of to show you the results are using Python (pandas, matplotlib, seaborn libraries) so here we go: (I’m skipping some data wrangling bits here..)
Glucose level distribution as a histogram
Glucose Level Time Series with rolling statistics
Honestly this data has me a bit worried because generally doctors, Peter Attia MD and the Levels program all suggest that fasting glucose should be less than 100/mg/dl…
Also, I wanted to correlate the glucose readings with the time of day, so we can do that eg with a heatmap:
Glucose Levels by Day, Hour in a Heatmap
This chart is really interesting to me because here a number of things stand out to me:
- I was in Austin from April 26th to April 28th for a conference, and those days I did a light workout in the morning, and had a light lunch. So clearly a lighter lunch, moving around in the afternoon leads to lower fasting glucose.
- I slept really badly Fri-28th-Sat 29th due to a late flight, and clearly a bad night sleep results in a bad fasting glucose.
- I generally workout in the afternoon – between 4PM and 6PM, and so higher readings there are not alarming to me.
- What strikes me as odd is the differences in the morning fasting glucose -say 5AM to 10AM – varies between 66 mg/dl to 120 mg/dl…The mean is still around 100 mg/dl (which is not great), but I’m surprised about the high variability.
- Since I do intermittent fasting (lunch is my first meal) I had generally thought that my morning glucose would be lower.
- There is a missing block on May 2nd as I switched the old sensor to the new one – as you have to do that every 10 days. The most painful thing was tearing the Levels patch off my hairy arms 🙂
All in all I’m very happy to have all this data from the CGM/Levels to explore, giving a lot of actionable intelligence – so I will try some life-style, diet modifications soon.