Is all fitness tracker data just an n=1 experiment?

Hi there, it’s Livnah. Back again for the Weekly Chart and today we’ll be looking at some cyclical health data with subject n=1 (me 🙂)

I’ve been wearing a fitness tracker and logging additional data into its app for eight years now. While I can easily see causations on a day-to-day basis — the earlier I eat dinner, the lower my resting heart rate will be during the night, and the earlier I go to bed, the longer my deep sleep phase is compared to light and REM sleep — I never looked beyond daily or weekly measurements.

Then last year I took a Datawrapper field trip to the Info+ conference in Edinburgh, where I saw Shirley Wu’s art piece “Though a patriarchy would privilege the changelessness of the sun over the inconstancy of the moon and you,” and I was inspired to also take a closer look into my health data that cyclically repeats. There are different measures like “breaths a minute,” average sleep, and body temperature, and I was surprised by how different the charts came out over time, even when not that much changed in my daily life.

In the following two charts you see changes in my basal body temperature over two three-month periods. I thought back over the past 8 years and picked one timespan around what I subjectively think was my best mental and physical health month (low stress, consistent bedtime, eating healthy, happy social interactions) and one timespan that was particularly… let’s say challenging (even though physically not much was different). I was very surprised by how different they look:

The first chart aligns closely with the textbook model of a female cycle, where basal body temperature changes in direct correlation with progesterone levels, rising in the second half of a cycle.

A fitness tracker offers more than just daily step counts. I was surprised to discover that the data I had collected over the past eight years extended beyond the daily comparisons that I had been using it for. I had never really looked for repeating patterns over longer periods. The few times when measuring my temperature with a physical thermometer, it was always in the higher 36.X°C or low 37.X°C, which in my mind was "the same." However, using a more sensitive measuring device with two decimal places and seeing the data displayed in a graph revealed much more.

I plan to continue the search for more patterns in my data and if I find something interesting might share more insights with my next Weekly Chart duties.

---

That's all for today! Next week's Weekly Chart will be from our designer Gustav.