On October 3rd 2011, at 13h46, I was collecting material for my master thesis about the influence of mobile devices on our lives and our social environments. More precisely, I was working on the Quantified Self movement, i.e. people fascinated by gathering and analyzing mundane events from their lives in order to understand themselves better and possibly share their findings with the world. Even though this tendency to log oneself is much older than they are, smart mobile devices have become huge enablers and the epicenter of most lifeloggers (recent lifelogging devices like the Jawbone Up, Nike Fuelband or Fitbit Flex pretty much rely on smartphones to operate). Smartphones are actually so effective they are gradually getting us involved with lifelogging too, whether we want it or not.
My first phone was only able to store 150 SMS, which was quite good at the time. I remember spending a few minutes every month removing most of them to make room for new ones. Nowadays, they are not only capable of storing tens of thousands of these messages, they can even help us search inside each and every one of them instantly. Which leads to uncomfortable situations, like stumbling upon an old evocative message sent to a then-girlfriend while looking for something else completely unrelated, except for one keyword taken out of context.
In the same vein, two years ago, millions of people realized they were being tracked at all time when it was discovered that most iPhones and some iPads had been recording their geolocation in a plain text file on the device itself↓. Of course, it’s not so much the device’s positions that matters but since we have grown accustomed to living with them at all time, these datas were actually about us. Behind each time/space coordinate available in this file was the certitude that I had been precisely there at this moment in the past. This is me at some point in my life, the first time I met this person I care about now, the last time I slept over at my old place, or this night on January 1st where I walked across the city for no particular reason.
In the trail of the iPhone tracking scandal, a great project was created to help people visualize this data and see what could be understood from it. Open Paths is the brainchild of Jer Thorp, an influential artist and designer in dataviz and interactive systems. The TED talk he gave on his work and this project is pretty good by the way.
What was nice with this debacle was the awareness that came with it. Yes, your device can track you, but so can a bunch of other actors: for example, european network operators are forced by the directive 2006/24/EC of the European Parliament↓ to store and provide law enforcement agency any personal data that they can have about any of their customers (except for the actual content of these communications) for a period of at least 6 months. But the data that they have on every one of us is what we provide them with, consciously or not. We have no idea we are giving them all those informations because they travel so seamlessly all the time, going from device to device, from servers to servers all other the globe. That’s the thing with seamlessness, it just works but you have no idea how. To see what this data says about us means making it visible, designing seams in the system. Intelligent, legible and beautiful seams.
Anyhow, at this point, I think we can define two types of lifelogging information↓. The first is what I’ve just talked about: traces of where we were, who we were with, what was the content of our communications, what I just had for breakfast, etc. Traces are entered automatically, without any action on the user’s behalf. In the case of network’s geolocation data, these traces are seamless, invisible, they hide below the surface, out of reach. In the case of communications such as mails and SMS, they are just lying there, still visible and intact. They stay precisely in the same state they were in when they were created, acting as the remaining witnesses of a discussion, an exchange of ideas that are probably not relevant anymore.
Marks, on the contrary, are what we leave on purpose. They are made to withstand the test of time (in internet talk, that’s saying something). They are usually much more significant and meaningful, the way a shrine on the side of the road in occidental societies suggests a fatal car accident at this precise location. Marks reveal human involvement while traces are just generated by-products of another activity. Incidentally, marks are generally of higher value to people and companies than traces. They are a result of engagement and attention directed toward something, they are the expression of an opinion made to stay.
That is not to say that this distinction is permanent. Traces can become marks. When playing for a long time on a game server, the Steam interface asks you if you want to add this particular server to your favorites right after leaving. In the process, they are effectively trying to turn simple traces (server history) into marks (bookmarks for later). I don’t think this mechanism would work with Google Search but it would be interesting to try: “you have just searched for naked ewoks gangbang C-3PO, would you like us to link this query to your profile and filter all your Google searches according to these terms for the rest of your life?”. An intriguing experiment, as well as an approach for addressing what is now called the filter bubble (essentially a side effect of seamlessness).
Amazon also has effective strategies when it comes to turning traces into marks: for example, a while after receiving a product that has few or no reviews, they send an email asking casually if everything’s fine with this item and if you wouldn’t mind rating and reviewing it. The email is very concise and straight to the point, which is interesting: a lot of websites use incentives like giving vouchers to get reviews, but here they are basically just asking their customers to spend time writing it. Therefore, when some of those customers do write reviews, their motivation is not extrinsic but intrinsic. It is done for the greater good, because these people know their knowledge on the subject has value to others.
The ratings by themselves also serve a purpose, one of enhancing the recommendation engine. If you want to get good personalized recommendations, you need to let Amazon know what you enjoyed. Sure, they can tap into the list of what you bought from them but it’s usually quite limited, and what about that CD someone gave you or that book you got at a book signing event? Foursquare puts it best on their blog, in a post detailing a recent update↓:
Remember: each time you check in, you’re teaching us about the restaurants, bars, and shops you like, so we can give you even smarter recommendations in Explore. And now that it takes just a split-second, it’s even easier to check in everywhere you go.
They may as well say they are losing money for the benefit of their customers. However, a leaked document detailing Foursquare’s strategy↓ revealed last week what this data collecting is really for :
In June, Foursquare will start offering the “post check-in units” on a cost-per-click basis. Foursquare will show ads to people that are relevant as soon as they check-in on the app.
Now this sounds much more balanced. I guess that’s a deal every Foursquare user has to carefully consider before checking-in next time. Thankfully, Foursquare allows you to export the raw data you entered on their plate-form, which means that theoretically you could switch to a competitor and go on your daily check-ins if you get tired of being served ads.
There is something fascinating about lifelogging. It might sound quite egocentric but it’s actually not. R. Buckminster Fuller puts it nicely in his book Guinea Pig B:
I am not being messianically motivated in undertaking this experiment, nor do I think I am someone very special and different from other humans. The design of all humans, like all else in Universe, transcends human comprehension of “how come” their mysterious, a priori, complexedly designed existence.
Buckminster Fuller is probably one of the most incredible human being of the XXth century. The experiment he is describing is about seeing the positive impact one person can have on the world around her in a lifetime. This hands-on approach is an integral part of the design process and recognizes failure as a necessary part of the exploration.
Meanwhile, my lifelogging experiments are much more modest: I am mainly concerned with finding ways to make the invisible visible, to give substance to time and to develop self-reflection through data and design. I am convinced we are moving toward an age of ambient self-quantification (self medication, personal coaching, individual recommendations, etc.), and it is vital to explore most of its facets before its there.
On October 3rd 2011, at 13h46, I entered the first log for this project on my smartphone. I used the app Tap Log, and later developed my own app, Log. I set up 9 categories that describe more or less objectively my days, and logged my life manually for the next 9 months. Basically, a year of preparing, writing and developing a design project for my master’s degree.
For some reason, in addition to the 9 categories I also logged what I ate but stopped after 7 months. What I got at the end was a huge CSV file with 4613 entries. Each entry indicates a new activity that I am starting, and it lasts until I switch to another one. Here is a list of these activities and the time I spent on each of them over those 261 days.
DORMIR (sleeping): 1980 hours DÉTENTE (relaxing): 1308 hours MÉMOIRE (work for the master thesis): 1098 hours MANGER (eating): 494 hours FRIENDS (time with friends): 451 hours BOULOT (freelance work): 250 hours HYGIENE (showers, shaving, etc.): 109 hours LABO (the hackerspace I started at my school): 63 hours ELSE (the rest) : 515 hours
It’s quite difficult to estimate how much it actually represents. 2000 hours of sleep seems like a lot, but over 260 days it is hardly 8 hours per day. Similarly, I am not sure how I feel about spending as much time working (MÉMOIRE + BOULOT) than relaxing. Activities are not that compartmentalized, especially in design. How should I log watching a documentary on typography with friends? For the same reason, I realized that shower time is actually one of the most useful part of my day for thinking about work and planning the day ahead. Another observation is that you don’t really “switch” activities. One task is the main focus but you can’t disconnect completely from the others. There is always one level of multi-tasking involved, mainly with adjacent activities. Which is why I used beziers and not rectangles in the visualization below.
When I think about this experience, I realize it is one of the most painful and tiring project I ever started. In the last few weeks, after 4000 or so manual entries, it got to a point where logging became a huge chore. Not so much because of the logging process (it was annoying but quite fast) but mainly because it got me to take a step back from life and lose part of its intuitive, relaxed and spontaneous ways. Every single action became somehow significant even when they were supposed to be really straightforward and anecdotal. I don’t think you can underline too much how valuable those meaningless moments are to our mental health.
Which ultimately begs the question, does measuring a phenomena change its nature? There is an interesting parallel to make with particle physics here, more specifically about the observer effect. In essence, it states that the simple action of looking at particles changes them in a way, and so what is observed is never in the state it was in when it wasn’t observed. Experimenting with lifelogging helped me see the major role this effect played on the way I functioned. Being aware of how much time I spent during the day doing something influenced what I wanted to do next. Also, when sleeping too much, I felt guilty about not logging more work. When logging 12-14 hours of work a day, I was under the impression I just did it for the logs. When skipping lunch, I ended up eating something a few hours later not because I was hungry but just to keep some consistency in my logs. At some point, I enjoyed going out really late with friends because it meant a rupture in the log and I liked what it said about me. Stating these reflections makes them seem pointless, but I believe they apply to our internet profiles to some degree. We have all become our own PR managers, carefully considering what the traces and marks one can find about ourselves say about who we are. Basically, SEO for our digital footprints.
As far as metaphors with particles go, I really like to think of Heisenberg’s Uncertainty Principle for this project. In simple terms, it states that the more you know about the position of a particle, the less precisely you can measure its momentum (and vice-versa). I’ve seen it confused with the observer effect more than once, but it’s not the same thing. The way I see it applied to lifelog would be something along those lines: the more we record real-time personal data, the more we miss the bigger picture and the less we understand in which direction we are going. Now I am not saying that it is deceptive to log some of our personal information. But there is a propensity in the lifelogging communities to believe that the more we log, the better off we will be. Also, there is clearly no “one size fits all” in this context.
Could it be that one of the defining difference between the human brain and a machine’s is its ability to forget, decay and die? We tend to see forgetting as a major failure of our brains, one that could be overcome with technology. But it’s probably one of its more advanced feature! We should cherish it, and use it as much as we can. I don’t want to stumble upon a chain of mails that my brain estimated needed to be removed from my memory. There has to be a reason for that, and most of the time there is. On the contrary, images, odors, sounds, or whatever data we gather should be designed as cues to trigger autobiographical memories of important people, places and events. We should not try to log memories that leave no place for reinterpretation.
In the movie Blade Runner, the replicant Roy Batty has this fantastic monolog when he is about to die:
I’ve seen things you people wouldn’t believe. Attack ships on fire off the shoulder of Orion. I watched c-beams glitter in the dark near the Tannhäuser Gate. All those moments will be lost in time, like tears in rain.
There is a distinctive humanlike tone behind these words, one that evokes our struggle with the passing of time and the value of traces we leave behind us. These traces can be anything, from raw data to stories and even memories inside other people’s head. Which reminds me of this surprising statistic by UNICEF↓ I heard about recently that basically states that as of 2013 over 40% of the word population will die with no records. With the current obsession with data, how long is it going to take before we get this percentage to fall below 20, 10 or even 1%? But most importantly, what kind of records should we keep?