Fitness and other tracking systems have the potential to measure some aspects of quality of life and possibly suggest changes that will improve future quality of life. I have worn a simple fitness tracker (a pedometer-style step-counting watch) for almost two years and have now combined the fitness tracker data with other sources – my real-life and online social activity – to look at how quality of life might be assessed.

The fitness tracker tells me my total steps and primitive measures of sleep. The real-life activities are social and work activities from my diary and photographs from my camera (which fill in gaps about where I was). The online activity is taken from the number of email and Twitter messages that I have sent per day. Some of the data is noisy and I have averaged over a week.

The take-home message is that physical activity improves quality of life. Twitter (or at least the way I have been using Twitter) has a negative impact. Doing more to be outside the house, even simply taking my activity to another place, would improve my fitness and happiness.

The fitness tracker I used is a Garmin Vivofit which replaces my watch and has a battery life of 1 to 2 years. I only have to remember to synchronise it (with a mobile phone or a laptop) every couple of weeks, by pressing the button beside the display. The only other interaction with the tracker is to move whenever the red lines (at the top of the display) show that I haven’t taken enough steps in the past hour. It is no different to wearing any other watch and displays the time and date. The Garmin estimates deep sleep, light sleep and awake time, as well as showing steps during the night.

More advanced fitness trackers can measure heart rate, galvanic skin response and include GPS position sensors, but almost all require charging every 1-5 days. A GPS-enabled tracker will probably need charging and synchronising every day. See this TechRadar review of fitness trackers for an idea of the range of features available.

The Garmin Vivofit may be a very suitable choice to use with an adolescent or adult like me who does not have the structure or patience to manage regular charging and synchronisation – just wear it as a watch and the data becomes almost magically available online or for download.

Pedometer function
My average daily step-count falls into one of three patterns of low, moderate and high levels of activity. These visually divide between 0–4,000 steps on 53% of all days (equating to around 0–3 km), 4,000–7,000 steps on 32% of days (3–5 km) and 7,000–15,000 steps on 15% of days (5–11 km). This means I probably did not leave my home on 195 days per year, walked somewhere within 2 km of home on 115 days per year and walked a round trip of 5 km or more on the remaining 55 days.
The weekly average of the number of steps is highly dependent on my social activities (or lack of them) and therefore very spiky data. The long-term trend shows a decline from just over 5,000 steps per day to well under 4,000 steps per day. The tracker has certainly not acted as a magic talisman to improve my physical activity levels and there are some visible life events: my walking was limited by a knee injury in January 2015, I was physically inactive while completing a book (The Painted Lorries of Pakistan) in May 2015 and again with an ankle injury in March 2016; my walking was increased with a trip to Dublin in November 2014, a long session taking photographs of sculpture in Cork in February 2015, a trip to Antrim in May 2015 and walking tour of Edinburgh in June 2016.

Calendar social activities

Combining my calendar adds a lot more information about the social activity on each day. I used data from a Google calendar and added in some missing activities that were obvious in my photograph collection. From these I can see the pattern of main activity and step count across time, with the main activity being:

• None (grey circle) – Most of my time, 156 days per year, I have no social activity, my step count is low (1,700 steps) and these are days that I spent alone;
• Shopping or walking in the city (black circle) – The next most frequent activity (80 days per year) is spent shopping or walking in the city alone or with family members, on average around 5,500 steps (4 km);
• Aspect (purple square) – The Asperger Syndrome Support Service usually takes me outdoors on about 40 days per year (just under once a week), often combined with a walk in the city and taking me to an average 4,100 steps;
• Lectures (green triangle) – Lectures on the far side of the city on about 26 days per year take me 7,700 steps (6 km), occasionally less if I get a lift in one or both directions;
  Travel (pink square) – Associated with the largest activity at 9,200 steps (7 km), usually walking while on holiday, about 20 days per year;
• Medical (blue square) – Regular appointments about 16 times a year are in an out-patient clinic across the city, on average 6,200 steps (5 km);
• Work (light green triangle) – Occasional meetings (9 per year) associated with books or research, usually somewhere conveniently nearby, an average of 5,500 steps (4 km);
• UCC (dark green triangle) – Occasional (unpaid) meetings associated with lecturing or marking are held in a building beyond the main campus on about 8 days per year, an average of 9,300 steps (7 km);
• Social (red square) – My least frequent activity, around 8 days per year, is social interaction that is not associated with work, averaging 5,700 steps (4 km)

The association between the main activity and steps is so strong that there is a clear increase in walking on Tuesdays (4,500 steps) and Thursdays (4,700 steps) when I lecture, and on Saturdays (4,500 steps) when I go shopping. I walk the least on Fridays (3,400 steps).

The boxplot puts the frequency of each activity type into perspective alongside the step count – travel is the activity associated with the highest step count, but it is an infrequent activity, and days with “None” are the most frequent (156 days per year) and associated with the lowest physical activity.

Time spent alone

The diary and fitness tracking data allows me to measure the duration of time alone, within the 195 days per year spent at home and 156 days per year with no social activity. Counting each sequence of days in which I did not leave my home and had no work or social activity (I did not see anyone else at home) there are 9 occasions lasting 1 week or longer in which I did not leave my house and nobody visited, a total of 15 weeks alone for a week or longer. The longest was 16 days, which was when my family went away for a trip.

Sleep tracking

The Garmin has a primitive sleep, categorised as “deep” and “light” sleep, as well as “awake” time that occurs during the intervals between deep and light sleep. Steps during the night are often picked up as awake time. The sleep times make visual sense as the time between the last sustained foot-step activity of one day and first sustained foot-step activity of the next. The plot on the left shows 7 hours and 15 minutes of sleep from 11:19 pm until 6:34 am (with a few steps at 3:10 am), while the plot on the right shows almost 12 hours of sleep including over half an hour reading before sleeping and a further four-and-a-half hours reading in the morning (after waking at 6:30 am) – the short walks to make tea and brush teeth weren’t enough to trigger the end of sleeep time.

The Garmin tracker identified my total sleep as averaging 8 hours 52 minutes per night – 4 hours 19 minutes of deep sleep, 4 hours 18 minutes of light sleep and 14 minutes awake during the sleep period per night. Deep and light sleep are negatively correlated and there has been a general trend towards less deep sleep, more light sleep ad more awake time.

Online social interaction

One interesting set of correlations are a positive association between emails sent and meetings with Aspect or UCC, and a negative correlation between emails sent and both travel and the total number of social activities. The correlations were even stronger with Twitter activity – my number of tweets sent was negatively associated with steps, sleep, travel and total number of social activities. This may be simply that I used Twitter (and less so email) when I had free time, or possibly that Twitter – or the way in which I use Twitter – has a negative impact of real-life social activity and sleep. It does seem to me that some Twitter interactions are deeply negative and that online social interaction reduces the capacity to be engaged with real life.

Conclusions

My physical activity levels are obviously too low and my physical fitness is poor. Social activity, work or photography seem to be the only motivators to get me out of my house. Temple Grandin (plainly much fitter than me) has said “I do 100 sit-ups a night, and I hate every one of them,” and I wonder what her motivation is to persist with something she hates. For myself I think I have to find more activities that require me to go out and to alter my relationship with online interaction.