In my previous post ‚An IR view on Lifelogging’, I outlined the use cases for lifelogging and challenges for the domain of information retrieval. In this post, I will give an overview of the different apps and devices I use to track my daily life.
I started tracking myself after organizing and joining the 2nd Workshop on Lifelong User Modeling. Our invited speaker, Cathal Gurrin, a very well known lifelogger from the Dublin City University, spoke about his experiences with lifelogging. He collected a visual lifelog containing over 12 million photos from the last 8 years. Judy Kay, professor at the University of Sydney, was wearing a Fitbit tracker and, if I recall it correctly, was taking part in a challenge counting steps between different university departments. Both got me motivated to start tracking myself. First of all, I had to decide what I would like to track and how the data should be collected and stored. Therefore, I defined some requirements for the tracking:
- Tracking should be automated as much as possible. At best, no user interaction should be necessary.
- The tracked data belongs to me, therefore, the services and tools must allow to access and control my data. Preferably, the data is under total control of the user.
- Tracking should not require a lot of extra devices. Device combining two or more functions are preferred.
Hardware-wise, the decision was obvious. With my smartphone, I already had a sensor platform at hand. My current iPhone 5s is still the main tracking device I am using. I extended the hardware collection with a Fitbit one tracker and recently an Autographer (I’ll explain why later on). Software-wise, the decision was not as easy as with the phone. The information I’m most interested in is definitely location information – preferably locations and the paths. For locations, I am using foursquare since March 2010 (now Swarm) and did around 12.000 check ins so far. But not only the target destination, also the way too it interests me. That’s how I discovered Moves. Moves allows you to track activities, walking, cycling and running. It counts steps and travelled distance, and also the GPS track. It is also connected to foursquare and allows you to link places to the according foursquare place. These two apps allow me to keep track of my daily movements – without significant effort.
Soon after started tracking my locations and movements, I felt the urge to collect more. Photos would be a nice addition to locations I thought. Not only knowing where I was, but also what I did. I decided to use flickr, wich has a decent auto-upload functionality. So, every photo I took was automatically stored online. Which eases the handling of photos a lot. But it still requires to take them manually. Which contradicts my first requirement. So I started looking for other possibilities to take photos automatically. Best possibilities I could find where the Autographer Camera and the Narrative Clip. Both are small devices that could be worn all day and take pictures every few seconds. I decided to give the Autograher (see my previous post about first experiences with the Autographer) a try for two reasons. It gives me full control over the pictures and it comes with some additional sensors. The problem is that it is an extra device, which needs to be charged, carried and turned on and off. So, I’m still not fully satisfied with this solution. Additionally, I’m also using the Everyday app. This apps’ only purpose is to take a picture of the user everyday, and compile a movie out of it – see this video for an example.
The current tracking was possible by using only my phone (battery problems included). Tracking my activities, e.g., steps, would be possible with the phone (the iPhone has an extra chip for tracking steps etc.), but I found it quite inaccurate as my phone usually lies somewhere near me, but is not always on me. So, step count was very unreliable and depended greatly on often I was wearing the phone in my pocket. I decided to use a Fitbit one tracker. A neat small device (and pretty durable as it survived several washing sessions) that allows to track, steps, conquered floors, how active I was and my sleep. I use the Fitbit since Juni 2013 and, except for a few days, it is a constant companion. The only thing missing is an activity detection component, to track my daily bike tours to work and back. Therefore, I use Endomondo, which sends the data back to my Fitbit app and fills the gap Fitbit leaves for activity tracking.
After locations, movements, pictures and activities, the last thing I’m currently tracking are my eating habits. I currently use MyFitnessPal and the Weightwatchers app. While MyFitnessPal actually fulfills my above defined requirements, I use the Weightwatchers app more frequently. Both apps connect to the FitBit app. MyFitnissPal sends summary of the used calories back to Fitbit. Weightwatchers takes the tracked activity from Fitbit and calculates how much more you are allowed to eat. I find the Weightwatchers app more motivating – first you have the points goal and then you get directly rewarded for activity. But MyFitnessPal is also a good app and also connects with Endomondo for instance.
In the lifelogging community, and especially the Quantified Self movement, one major argument for tracking ones life is that tracking yourself helps to live healthier and change bad habits. I actually started tracking just to have a personal lifelog. Data I can browse to remember what I did when for example. But I actually have to admit that it changed my behavior. I try to bike or walk as much as possible to reach my Fitbit goals (10.000 steps and 30 min of active minutes per day) and to get extra Weightwatchers points. And I changed my eating behavior. The small ‚snacks‘ during the day are history since I know for how many calories they are responsible.
Overall, the tracking for most parts of my daily life works pretty much automated or only requires a few actions. One challenge is still the data itself. Everything, except the autographer photos, are stored on the servers of the respective app. For most apps, there are APIs allowing to get the data, but it is still manual work to collect everything and to create a comprehensive personal lifelog.