The term “Quantified Self” was coined by Wired Magazine Editors Gary Wolf and Kevin Kelly in 2007. Wolf’s 2010 New York Times Magazine cover article is still seen as the defining “manifesto” of the Quantified Self movement. According to Kelly and Wolf’s website, http://quantifiedself.com, “Quantified Self is a collaboration of users and tool makers who share an interest in gaining self knowledge through self-tracking”. The typical examples are devices like Fitbits, or the the daily step data measured by your watch.
Part of me always recoiled at the idea of defining myself in terms of quantifiable data points — there’s something cold about it, something that makes me sure that the nuances of human existence don’t fit into a spreadsheet. I was never sure that counting steps gave me much in the way of “self knowledge” that I didn’t have before. Even if I was able to correlate more exercise or increased sleep time with a better mood — did I just learn anything that popular wisdom or a decent book couldn’t have told me?
On the other hand, self knowledge itself sounds like an important goal, and a very old problem. Somewhere behind all the other problems — building relationships, a life course, finding meaning in the things we do — the problem of self knowledge comes up. Understanding and being true to your own nature, and being able to change when the old stories you tell yourself stop working. I wasn’t sure that tracking data points could help me with that, but this is the story of why I might have been wrong.
QS usually emphasises the quantifying of data which can be measured automatically, especially by wearable sensors. Movement, sleep patterns, fitness tracking, sensors that measure biological signals to measure health. In this way it is closely related to Lifelogging, which is more closely associated with wearable cameras. Steve Mann and Cathal Gurrin are famous exponents of this movement, often wearing cameras on a continual basis. Steve Mann has coined the term “sousveillance” (observation from below) in opposition to “surveillance” (observation from above) to describe the use of recording technology by participants in an activity, as opposed to recording by devices physically or hierarchically “above” the participants. Microsoft researcher Gordon Bell’s 2009 book “Total recall” documents his attempts to log and organise all the information he can related to his daily life, from health data to family photos (I can only assume that lots of grey tiled windows were involved).
Closely tied to Quantified Self is a rising interest in data visualisation. After all, you’d rather see your data in the form of pretty graphs than as rows and columns. An example of data visualisation using a Quantified Self approach is the work of Nick Feltron, whose website http://feltron.com collects Feltron’s “annual reports” which document aspects of his personal life each year. These highly aesthetic graphs visualise information such as Feltron’s sleep patterns, eating habits, communication methods and personal relationships. This kind of recording of potentially subjective data is more appealing to me than the more physical applications of Quantified Self. With Feltron’s apps “Reporter” and “Daytum”, you can even try tracking yourself. Reporter allows people to specify a number of questions to be asked at random intervals throughout the day, and generates visualisations based on this data.
Unexplored areas of Quantified Self
When I think about “knowing myself”, I think about ideas like passions and authenticity, not so much about calories, kilometres and body-mass index. I’m willing to accept that this is a personal preference. My perception of Quantified Self was always that it focussed more on the latter. I started to realise, though, that this didn’t have to be the case.
I may not be an exercise fanatic, but I make up for that with other strange interests. One of those is my dream diary. I’ve kept up this habit on and off since early high school. A few years ago, I was curious enough to poke around and see what the options were for recording my dreams online. I could never carry my physical diaries around with me everywhere I moved, but there was something cold and businesslike about firing up Microsoft Word to record a dream.
There were a few websites out there: Dreamboard, dreamjournal.net, even an offering from Romania called “Dreamophone”. I was expecting simple blogs, with the ability to read dreams that other people had posted. Something about what I found though surprised me. What I found was data. Graphs, statistics, little word clouds that reflected what individual dreamers or the users of the site were dreaming about. At first I was turned off by these features — they seemed like an unwelcome intrusion of business metrics into the creative and fantastical world of dreams. When I started using them though, I started to change my mind. I was learning things.
The quantitative study of dreams
A segue into history. The quantitative study of dreams began in the 1940s, originating with the work of Calvin S. Hall and later Robert Van De Castle. The two worked together to create a system of content analysis of dreams known as the Hall/Van de Castle coding system. Essentially they advocated creating a set of metadata about a given dream (and sets of dreams) and analysing the resulting data sets. This dream “metadata” could include such variables as:
• Characters (animals, men and women, friends, strangers)
• Objects (chairs, cars, streets, body parts)
• Settings (indoors vs. outdoors, familiar vs. unfamiliar, specific places)
• Activities (thinking, talking, running)
• Emotions (happy, sad, embarrassed)
• Social interactions (aggression, friendliness, sexuality)
According to Hall and Van de Castle, even this surface content of dreams can give interesting insights into the dreamer, especially when analysed over time. (They were usually more interested in spreadsheets than pretty graphs though)
First attempts
After a few months of using the available websites though, I had mostly exhausted the Quantified Self aspects they had to offer (for example, most data was cumulative, so eventually the latest data barely registered amongst the weight of preceding weeks). So I gathered up a pile of my high school dream diaries, and tried to visualise the information in them based on the Hall/Van de Castle system. The result was a fairly simple graph using Processing, visualising the people, places, actions, things, themes and emotions that most often occurred in my dreams in a given year. The more I dreamt about something, the bigger the bubble. The slider at the bottom allowed me to slide through the years, from 1997 to 2002.
It’s still hard to describe the enjoyment that I got out of this. Every state of the graph is like an portrait of myself at that age, represented in an abstract painting. I see concerns coming and going, high school friends and crushes, important places and objects, and how I was feeling. As the slider moves, narratives come and go.
I realise how much certain people meant to me, or perhaps how much they represented for me. I see patterns emerge, and desires and fears as they grow and shrink.
Exactly the kinds of things I see? Maybe that’s not interesting to anyone except me, or maybe I don’t feel up to explaining it yet. I didn’t record nearly as many dreams in high school as I imagined, so the data isn’t as great as I thought it could be. So all the detail might be the topic for another post. But sufficed to say, even these basic experiments were quite powerful for me. They’ve given me a new perspective on Quantified Self, especially about visualising subjective and psychological information. Perhaps there’s self knowledge that I can get from my data after all.
