It took me a while to put together but I’ve just released by new video and spreadsheet tutorial looking at exercise vs happiness.
There’s a blog post here:
Or you can just watch the video here:
TLDR: Exercise is good for mood!
I want to cover off a few more topics before I tie up this series, but it would be encouraging to see other people trying this analysis for themselves. Are you likely to try this or some variation on the this theme?
This is a really superb walk through of one person’s approach to answering a personal question with self-tracking data. I don’t use these particular tools so I can’t imitate the project exactly, but I just want to acknowledge and say thank you to @michaelforrest for taking the time to present this approach with such clarity and detail. If you are comparing two sets of time series observations there are some methods here that are useful to see.
That’s amazing I was looking for the facts, about the impact of exercise on mental health.
Maybe? What if there is another factor that both prevents exercise and depresses mood (e.g. a stressful schedule); taking that into account could in some cases reverse the relationship between exercise and mood.
Either way, great walk-through!
A glib conclusion, I will admit 
My thesis around happiness data is, though, that it’s impossible to get a particularly scientific result when there are so many possible factors influencing mood.
Instead, I look at it as a way to give myself evidence to support continuing behaviours that I already know are beneficial. A chart like this sticks in the mind.
I’d definitely be interested to see how others interpret their data. I think there will be interesting psychological differences between when different people choose to call their analysis “complete”.
What if the chart had ended up showing less happiness after exercising? Could happen e.g. if you exercise more on weekdays, but are happier on weekends… Would that have reduced your willingness to exercise?
That’s the question isn’t it. I suspect I would have continued digging into the data until I found a more satisfying conclusion but I would have to concede something if I wasn’t able to reconcile my expectations with the data. I don’t think I would have gone straight to not exercising, but perhaps I would have tried to find a way to have a better exercise experience. Later I could compare the data before and after the change to see if my change had a positive effect. Rinse and repeat.
That makes sense… Looking forward to your next tutorial on A/B testing 