Lessons from a Nomad's Apple Watch Data

In the last year, I've lived in 13 cities, usually for a little less than a month each. I wanted to know which attributes of a city helped me stay active & healthy so that I can select for those attributes for my travels in 2022. To figure this out, I exported 365 day's worth of my Apple Watch data into a CSV and cross-listed it with my travel history. I played around with some PivotTables and tried to learn a few things about my body/activity habits.

Disclaimer that I'm not a data scientist and that this 'analysis' is about as basic as you can get. I initially did this for my own knowledge and then a few friends encouraged me to share. Ultimately, fitness should not be about obsessing over metrics. What's most important is to move and feel good in your own body. I'm just Type A and was bored one day, and now here we are.

This piece is divided into three parts. The first section is what I learned from my data. The second section is a step-by-step guide on how to recreate what I did, for anyone who is interested in learning this information about themselves. The last section is a wish-list for the tools I need to make this analysis more comprehensive.

The data + what I learned

Quick observations:

• I was most active in Portland. It's where I exercised the most & burned the most active energy of all 13 cities on this list. I also had the lowest resting heart rate in Portland. Generally, a lower RHR implies more efficient heart function and better cardiovascular fitness.
• My highest RHR was in Hawaii, where I was not sleeping very well. I was waking up at 3AM for 9 EST meetings so it makes sense that the place where I was sleeping poorly led to the highest RHR. I also had higher-than-my-average RHRs in San Diego and Los Angeles - both places where my sleeping environments weren't ideal. In LA, I was sleeping on the floor. In San Diego, I was living in a house with 34 others.
• The three cities where I was most active (Portland, Jackson, Austin) are also the three cities where I biked to and from the gym (as opposed to walking, driving, or taking public transit).
• My lowest average exercise times were in Los Angeles and when I was in Orlando for a conference. Makes sense - I was busy with meetings and events all day, and was generally very stationary. Only had time for the treadmill at night.
• The two cities where I climbed the most flights of stairs were both mountain communities - Appalachia (I was staying near Boone in North Carolina) and Jackson, Wyoming. That makes sense - in both places I was taking advantage of the beautiful hiking around me. The next two highest (Jim Thorpe & Hawaii) also both involved some hiking. Cool to see that the watch detects ascension successfully.
• I was surprised to see that places where I had standing desks (Philly, No. Virginia) didn't always correlate with high standing time scores. Maybe the watch isn't great at detecting standing, or maybe my attempts to improvise standing desks out of stacked chairs/high ledges were adequate-enough that the standing times weren't impacted by location.
• Philly & NYC had almost identical average exercise minutes, but I was burning way more active calories in Philly than NYC. This is probably because I'm able to play pickup basketball much more frequently in Philly. Basketball typically burns more calories/hour than any other activity that I regularly log.
• My number of housemates didn't seem to strongly impact my activity levels. There was no correlation between my activity levels and places I lived with lots of folks (San Diego, Hawaii, Jim Thorpe) compared to places where I had more normal roommate situations. That's good - that data indicates that I'm still able to prioritize regularly activity without worrying about accountability or distraction.
• Over a third of my 15 most active days (measured below by active calories burned) were in No. Virginia where I had easy access to a large network of bike trails.

How I'll apply these lessons moving forward:

Future neighborhoods I explore need to have housing with easy access to at least two of the following:

• Bike trails
• Basketball courts
• Mountains with good hiking.
• An olympic pool/ocean for swimming
• Ultimate Frisbee pick up games.

I went on runs and to the gym in all 13 cities, but I only regularly over-achieved in cities that had at least two of those five elements. The more of those elements a city had (Portland had 4/5, Austin, Wyoming, & Hawaii each had 3/5), the more active I was.

If I do choose to live in a car-centric city, I need to ensure that I live within a walkable oasis within that metropolis. Orlando, LA, and Austin are all very car-centric, but at least in Austin I lived near the river and downtown area, which enabled me to rely on my bike to keep my body moving.

Community living may not have impacted my activity metrics, but living with friends definitely helped my mental health stay in great shape. If I can balance the RHR + sleep lessons with my communal nomad travels, I'll be a happy camper.

Finally, I'm no statistician, like I mentioned at the top. If there are other analyses I could have done with this data that I should have considered, please let me know. PivotTables and excel filtering were the only tools I thought to use.

How to analyze your own Apple Watch data

Apple already provides basic dashboards for viewing your data. Here are a few screenshots below from the Activity app:

These provide solid basic insights and visuals. However, Apple doesn't allow you to easily export a CSV. If you're a neurotic nomad like me or you otherwise want to go deeper to manipulate the data yourself, these are the steps I followed:

1. Download the Health Export app from the Apple store
2. Export Aggregated Data and choose dates for the past year. Only choose the fields that you care to analyze. I chose all of them except 'swimming stroke count' and 'distance downhill skiing'.
3. Open the table in Excel and add a column for the cities you've lived in (I searched through my email inbox for Amtrak & plane tickets from the last year to get my dates correct).
4. Create a Pivot Table with "Cities" in the 'rows' box and start experimenting by dragging different fields into the "Values" box.

What I want from the wearables of the future:

My health data that I analyzed from above is far from complete. Here are features I would like to see included in future devices:

• Sophisticated blood sugar tracking. There's no nutrition data to pull into this analysis, which is a HUGE factor when thinking about holistic health. I would love to be able to better track my blood glucose levels. There are a few companies working on this (Levels Health, Supersapiens) but I have so far been unable to get access. Ideally, Apple could just integrate this into their existing product. They are rumored to be doing this, but FDA regulation may be a hurdle that they would rather avoid.
• Better auto-detection. The Apple Watch (I have the Series 5) has to be told when I'm exercising, and how I'm exercising. This means that sometimes it doesn't detect light brisk walks or shorter bike rides if I forget to record. This makes the data imperfect.
• An auto-save feature (and increased battery life). My watch died during three different longer bike rides and thus those workouts weren't recorded. It happened during other types of workouts too, especially on days where I forgot to charge my watch or plugged it in wrong. Would love to see Apple at least save the progress made on a workout so far if the battery is going to fail, instead of erasing the workout completely. Also, if watches could charge without coming off my wrist, then maybe I could start analyzing my sleep, too.
• Improved accuracy. Any fitness professional is skeptical about the accuracy of Apple's calorie tracking data. I'm not sure how it calculates my number of calories burned during exercise, but I suspect it's some sort of formula based off of what kind of workout I say I'm logging. That doesn't really make sense. If I'm playing basketball but not really trying, then the watch might say I'm burning more calories than a swimming session in which I'm going all out. In a perfect world, I shouldn't need to say what kind of work out I'm doing (or that I'm working out at all) - the watch should just be able to understand caloric burn based on my movement, heart rate, and other biometric readings that it gathers.

Concluding thoughts

The first city I'm visiting in 2022 is CDMX. Psyched to be back in Mexico (where I went to Kindergarten!), but I'm worried about my ability to stay active in such an urban environment, based off of these lessons. I haven't booked lodging yet, so I'll be looking for neighborhoods with the features I described above.

Ultimately, this was a fun exercise, but was rather one-dimensional. Activity metrics are just one piece of health. Sleep, diet, environmental factors, relationships, and so much more influence health & happiness. Trying to decide where to go based off of just one set of metrics is too limiting.

I have since found out about a service called Nomad List, which enables nomads like myself to filter cities based off hundreds of factors, including weather, safety, cost, and much more. Highly recommend using it if you're trying to decide where to go to next!