For the last week of Rest of You, I tried to automate all of the trackers that I made in the past seven weeks, and feed them into a locally hosted visualization. I didn’t get quite that far, but am getting there:
I liked the viz I created last week, so my obvious first step was to automate the python scripts that retrieved the data. The productivity scores are requested from the RescueTime API, and my Chrome history is just a JSON converted from a SQL file that Chrome locally maintains. Both python scripts update their datasets hourly. The JSONs are linked to a local d3.js visualization, so I can refresh any time to see an update. The design could use some cleaning up, but the data is there.
The next step was to automate the entrainment process as well. Rather than having to monitor my own mind-wandering, then muster the self-discipline to open up my six-minute entrainment meditation, a python script checks my productivity levels and opens the webpage if they dip below 50%:
If I “meditate”, the event gets encoded as a bright teal line in the viz, so I can see how the entrainment affects my behavior. I was apparently pretty productive this week, and didn’t trigger the entrainment script at all, so I had to waste some time this morning to see it in action:
Finally, I also got around to hooking up my python keylogger to IBM Watson’s Tone Analyzer API. I was hoping to feed its analysis into my visualization as well, to see if mood was correlated to mind-wandering, but the results were a little questionable:
Moving forward, I’d also like to integrate Affectiva’s attention, engagement, and valence variables into the viz—I just need to figure out how to open the locally-hosted demo in the background, so it captures a candid moment for a few minutes every hour. I would also like to figure out how to automatically count the number of windows and browser tabs I have open, as I’m pretty confident these are reliable markers for stress and mind-wandering. Finally, I also plan to create a questionnaire that prompts me every hour, to have corroborate the correlations of all my tracked data.