At the beginning of the year I proposed to track my mood via Moodscope (subjective measure) while body blogging (physiological measure), what follows is my initial impressions so far in using these technologies before presenting my conclusions at the Quantified Self Conference in Amsterdam in November.
Impressions of Moodscope
I am aiming to combine these two ways of recording my activity to see if my physiology can tell me something about my mood states or my mood states about my physiology. I have kept a diary alongside Moodscope, where I noted down my activities and feelings throughout the day. This was important as my average score was, as it had been before at 56%, but I found that my diary entries did not reflect that I was not happy, but simply that I had a normal working day. I therefore started to keep a separate note of my pos/neg ratio which revealed that I had a 3/1 ratio most of the time. The ratio is calculated by adding up all your positive and negative score separately and dividing each by 10 (the amount of pos/neg words) and is the original way the PANAS would be evaluated. This is important as the original PANAS suggests that one should be aiming for 3 positive to every 1 negative emotion experienced.
A second aspect I noted is that I had a couple of days where my Moodscope score fell to 30%, which is, even by my standards, very low. Looking at this in more detail it would reveal that I was sick on both occasions but by no means in a bad mood. Again when I looked at my pos/neg ratio I found that my score reflected a 1/1 ratio rather than the 3/1 ratio. This would mean that my positive scores had decreased significantly whilst my negative scores had stayed the same. This is important as I was not experiencing any negative feelings, I was simply not ‘active’ or ‘alert’ or ‘interested’, suggesting that my levels of physical activity have a great influence on my score.
The use of non-emotional words such as ‘active’ has previously been mentioned as a critique on the PANAS itself and appears to influence Moodscope in turn (e.g. Smith, 2008). On days where I felt active and alert my score was higher than on those where I felt low on energy and drained. It also confirmed an observation I had made whilst taking the test at different times of the day, i.e. I’m generally more alert in the mornings, after my first coffee that is, than in the afternoons which is reflected in my scores. I wonder therefore if Moodscope reflects to a great deal levels of activity, rather than overall mood (although one could argue that these are related to each other).
Impressions of Body Blogging
Measuring my heart rate for 8 hours (9am-5pm) on a day to day basis was a strange experience and I never got fully used to wearing the monitor as it needs to be worn fairly tight in order to give accurate recordings. I therefore never forgot that I was wearing it. With this came a constant awareness that someone out there was watching me that I could not shake off. I found this particularly strange at the beginning where I felt that I could not move from my desk at home unless it was for legitimate reason, such as lunch or the loo. After all I was meant to be putting in 8 hours studying at home. Although this aspect did not completely disappear it got a little better with time.
Sometimes I would start to doubt that anyone was watching, after all it’s not really something you talk about all day, but than a ‘unusual’ pattern appeared and Kiel would ask me what I had been up to that day, confirming that he did indeed notice bigger changes in my overall pattern. This ability to detect changes in my daily routine reflects Kiel’s great knowledge of the Bodybloggin system and the heart rate patterns produced by this which he gained whilst monitoring his own heart rate for over a year. Nevertheless, he reinforced with this my feeling of being watched.
There were other interesting reactions that I, as well as Kiel noticed. Since we did not have time to build up a separate system for my blogging I used Kiel’s Twitter account which is also connected to the system for the Physiological Computing site that changes the page’s colour depending on the heart rate. What is interesting about this is that my heart rate at rest (e.g. sitting at a desk) is about 20bpm lower than Kiel’s. This had two main implications. On twitter it meant that Kiel was in serious trouble as his heart rate had dropped from an average 70/80bpm to an average 45/55bpm. The Physiological Computing website, in turn, indicated that Kiel was asleep for 3 weeks, after which we managed to adjust the site to my heart rate and the colours reflected more accurate my activity levels. Interestingly no one seemed to have noticed, at least no one enquired to Kiel about the changes or checked that he was ok.
Similar to this was the reaction of my mother (who works professionally with learning platforms and communication technology, such as twitter, at an open university). When I first mentioned this project to her, her thoughts immediately turned to ethical considerations of a nation being watched, and monitored in new ways. Nonetheless, I invited her to follow my blog on twitter to receive updates about my heart rate every 30min. I received a reply saying that she would most certainly not sign up to this. I was surprised by this reaction but thought that perhaps she did not want to keep such a close eye on me as it could create worry in her should my heart rate show unexpected patterns. Out of curiosity I asked her about this, and she replied that she would simply find it ‘annoying’ and ‘disruptive’ to receive my feeds every 30min!
As surprised as I was at first by this reply I think it raises an interesting question, the assumption that other people would be interested to follow another person’s heart rate. Although I have told other people about what I’m doing I’m not aware that anyone has ever taken the trouble to look at or follow my heart rate changes on twitter. It would also reflect the lack of concern about Kiel’s heart rate “changes”.
This brings us also back to the question of how to make this data meaningful to a wider audience. After all, an average heart rate becomes meaningless if you lack the knowledge to interpret it. I myself feel that my daily scores barely inform me about my activity levels on that day. The most variation I can detect is the amount of housework I did that day or if I was at home or the office. My 30 min average would probably tell me even less, unless I had a good idea about heart rate patterns in general, and let’s be honest not many of us do not. Perhaps it is more patterns that form over weeks or months that will provide us with interesting results about aspects such as exercise or diet.
I can detect from my daily heart rate recordings the effects of coffee on an empty stomach, although this hasn’t changed my habit to start the day with a good cuppa. I’m therefore keen to couple the data with my Moodscope scores to see what they tell me about my daily mood and activities, and to see how I can get meaning out of my recordings that tell me a little more about my daily routines than not to drink coffee on an empty stomach. I will be presenting these results at this years Quantified Self Conference in Amsterdam in November.
Smith, R. (2008). The Long Slide to Happiness. [Article]. Journal of Philosophy of Education, 42(3/4), 559-573. doi: 10.1111/j.1467-9752.2008.00650.x
Pingback: Physiological Computing : Lifestreams, body blogging and sousveillance