I noticed that my most productive years were those where my computer time was divided into 40 minute chunks for technical reasons. I almost posted about just that, then I remembered that I had always been thinking about various other patterns over those same time periods, so I made myself a survey and filled it out for each period.

I only have 5 datapoints (each period is roughly 2 years long, other than the last two, which are closer to 3; I probably should have broken those into 3 periods), each with 8 variables. I gave each a value between 1 and 5, other than one boolean (the afore-mentioned forced pomodoro-like situation).

Without running any formal analysis, and just going by my impressions when copying the numbers, I found that productivity, in-person socializing, and overall mood are all positively correlated. Amount of sun was surprisingly correlated with nothing (other than air quality, because they're both easier to get outside). My two worst phases, in terms of productivity and mood, are the two phases where I did not have the forced pseudo-pomodoros. Exercise and diet appear to be weakly correlated with productivity/mood, but I also suspect I'm starting to show symptoms of diabetes and am getting less tolerant of foods that had no noticeable negative effects in the first phase, and to a lesser extent, the second. And finally, ease of internet access was inversely correlated to productivity/mood.

This tells me that, the next time my computer situation resets, I should go for the forced reboot situation, and avoid having internet within 10 meters of everything else I need (so not at home; go to a Starbucks when necessary). Optimizing the others is not quite so straightforward, but they can certainly inform future decisions.