With another Quantified Self Toronto meetup in a few weeks and a conversation with fellow self-trackers, it's time for me to think about time again.

I've been fixing bugs and adding small pieces of functionality to Quantified Awesome, and I spent some time improving the integration with Emacs. Now I can type ! to clock in on a task and update Quantified Awesome. Completing the task clocks me out in Emacs and updates Beeminder if appropriate. (I don't update Quantified Awesome when finishing a task, because I just clock into the next activity.) This allows me to take advantage of Org's clock reports for project and task-level time, at least for discretionary projects that involve my computer. I'm not going to get full coverage, but that's what Quantified Awesome's web interface is for. It takes very little effort to track things now, if I'm working off my to-do list. Even if I'm not, it still takes just a few taps on my phone to switch activities.

Most of my data is still medium-level, since I'm still getting the hang of sorting out my time in Emacs. Looking at data from 2014 so far, dropping partial weeks, and doing the analysis on April 14 (which is when I'm drafting this), here's what I've been finding.

• I sleep a little more than I used to: an average of 8.9 hours a day, or 37% of the time. This is up from 8.3 hours last year.
• It takes me about an hour to get ready in the mornings. If I have a quick breakfast instead of having rice and fried egg, I can get out the door in 30-45 minutes.
• It takes me 50-60 minutes to get downtown, whether this is by transit or bicycle. Commuting takes 3% of my time.

Little surprises:

• I've spent almost twice as much time on business building or discretionary productive activities (19%) as I have earning (11%) – good to see decisions in action!
• I've spent more time drawing than writing this year (5% vs 3%). Next to writing, Emacs is the productive discretionary activity I spend most of my time on (2%).
• I've spent 10% of my time this year on connecting with people, a surprisingly high number for me. E-mail takes 1% of my overall time.
• It turns out that yes, coding and drawing are negatively correlated (-0.63 considering all coding-related activities). But writing and drawing are positively correlated (0.44), which makes sense – I draw, and then I write a blog post to glue sketches together and give context. Earning is slightly negatively correlated with building business/skills (-0.15), but connecting is even more negatively correlated with time spent building business/skills (-0.35). So it's probably not that consulting takes me away from building skills. Sleep is slightly negatively correlated with all records related to socializing (-0.14), but strongly negatively correlated with productive discretionary activities (-0.55). Hmm. Something to tinker with.

Some things I'm learning from tracking time on specific tasks:

• Outlining doubles the time I take to write (and drops me from about ~30wpm to about 9wpm), but I feel that it makes things more structured.
• Drawing takes longer too, but it makes blog posts more interesting.
• Trying to dictate posts takes me way more time than outlining or typing it, since I'm not as used to organizing my thoughts that way.
• Encoding litter box data takes me about a minute per data point. So spending a lot of time trying to figure out computer vision and image processing in order to partially automate the process doesn't strictly make sense, but I'm doing it out of curiosity.
• I generally overestimate the time I need for programming-related tasks, which is surprising. That could just be me padding my estimates to account for distractions or to make myself feel great, though.
• I generally underestimate the time I need to write, especially if I'm figuring things out along the way.

This post took me 1:20 to draft (including data analysis), although to be fair, part of that involved a detour checking electricity use for an unrelated question. =)