As promised, posting today the slides of my “100 Days of Summer” talk that I presented at the 2013 Quantified Self conference in San Francisco last Thursday (October 10). In this presentation, I shared most important and interesting results of the first phase of Measured Me project.
The main objective of the first phase was to test 8 metrics that I chose to quantify my Being Well (current state of body, mind and psyche) and Existential Well-being (happiness and life satisfaction), by tracking them continuously for 100 days, several times a day. At the end of the summer, I looked at the trends, correlations, and other patterns and properties to evaluate “trackability” of each metric and their potential for describing and predicting my everyday life. My conclusion was that these metrics are relatively well suited for the purposes of quantifying and capturing my everyday life, and I will continue using them (although after some revisions and improvements).
The presentation was given as a part of the “Show and Tell” series. At the end of the talk, there were some interesting questions from the audience:
Gary Wolf (@agaricus), co-founder of Quantified Self movement, was interested in my approach to quantify emotional stability and mood swings on a daily basis. I have posted briefly about these metrics before, and will post in more detail soon
Rajeev Gupta, aka Toolmaker of QuantifiedSelf.com, asked if tracking my well-being indicators resulted in any changes. This is a very interesting and legitimate question: often, the process of observation itself affects the objects and subjects that are being observed (Hawthorne effect). My understanding is that the indicators of-well-being that I have been tracking so far are relatively uncontrollable, and have not been affected during the tracking process. In other words, just by noticing that I have been unhappy or tired in the past several days won’t necessarily make me more happy or energetic today or tomorrow. I do, however, expect issues with self-observing bias when I start tracking the drivers of well-being, like duration of sleep, time spent on certain activities, etc. I may start modifying my current behavior based on the observations from previous days; there are, however, some things I can do to control that, and I will be talking about it in my later posts
there were a couple of questions about statistical properties of my tracking data, including whether the emotions intensity and positivity scores were normally distributed. Before analyzing data, I did use a statistical technique called Shapiro-Wilk test to analyze distributions of each metric. The results were mixed, with some metrics showing normal distributions, and others being non-parametric. At the end, I decided to use non-parametric methods like Spearman’s rank correlations and mean rank tests to analyze the patterns in my data. I will be also making the data available for download on my blog soon; in the meantime, you can drop me a line and I will e-mail you the data set
I am currently preparing for the next phase of the project. The Quantified Winter phase will last four months (November – February), and will focus on improving some metrics (including physical health and cognitive tests) and testing new metrics (sleep quality, productivity, creativity, social capital). Stay tuned!