It all started with the realization that I could export data from my health app. If the data is there and I can mess with it, the temptation is too great not to do so. And if I can turn it into some interesting visuals, all the better.
During a few months in particular, the correlation was much stronger. Compare the R² in January to April, for instance. The R² should be somewhat representative of the proportion of variance in steps explained by the temperature.
So, in January, the temperature was much more predictive of my walking distance than in April. This makes sense - when temperatures had more fluctuation, my walking distance did as well.
It's interesting to look at the results of the models on a month-to-month basis. January's model makes sense, intuitively. I'd walk roughly 3x as much at 72°F as at -15°F. Compare that to March's model, which predicts that an 87°F dip in temperature from 72°F to -15°F would only cut down my walking distance by about 3,000 steps. I like walking, but not that much.
Unusual or unseasonable changes in weather led to outsized changes in walking distance. Essentially, a nice day in the midst of many cold ones motivated me to get outside more. The corollary: a cold day in the middle of nice ones meant less walking.
1. My walk to work in the morning was when I logged quite a few steps - 1,138 on average between 8am and 9am.
2. Usually, I walk across my office's campus to eat lunch. On average, that journey between 11am and 12pm nets me 716 steps.
3. Between 4pm and 6pm I usually take about 2,633 steps. Why so many more than on my way to work, though? Well, I often run some errands or grab groceries on the way.
4. I usually work out in the 7-9pm range, and the steps to go do so average about 1,308.
If this has piqued your interest, none of this is too difficult to do. You need your phone, a laptop, and a tool like Tableau, PowerBI, or Excel.