As I mentioned in my post about my personal wellness journey, I rowed in college and somewhat recently joined a rowing gym called Health House. For Christmas last year, I got an Apple Watch for the purposes of tracking calories while exercising. I’ve now had a few months of data before and after the macronutrient tracking initiative I started on March 15th of this year, so it’s time to examine the data to see what I can find. I’ve focused the most on hitting my daily protein goal, as you can see here, but I’m tracking protein, fiber, carbs, and fat intake.
Naturally, I created a Tableau dashboard for presenting my exercise data. However, before I get there, I’ll describe what I did to get the data – it’s a repeatable process, but the data had to be massaged. First, I grabbed data about my workouts from the MindBodyOnline website, which Health House uses to manage classes and schedules. From that dataset, I have the date of the workout, the teacher, the class, etc. I simply pasted that data into Excel. Relatively easy.
The harder part was dealing with the activity data from my watch / phone. The Apple Health phone app allows users to export data, but it comes in a format that the majority of people have no idea how to make meaningful – XML. Programming in R or Python or another language could change that into something I could use, so I looked for R code since I’ve had some experience with R. Some quick Googling turned up a website with instructions on how to convert the data. It didn’t take long to convert the data into a .csv.
However, the data is not aggregated; Apple Health records data at seemingly different intervals depending on which metric is being recorded. Those intervals are short. One second. Five seconds. Whatever. I needed the data aggregated to a day level to match my MindBodyOnline data, so I brought the .csv into Tableau and started aggregating. I could do this in R too, but I was also exploring the data. Tableau makes it much easier to explore data.
Again, before long, I had the data I wanted at the level of detail I wanted, and I exported that out to Excel. I could then join the two datasets together – one dataset with a row of data for each workout, and another dataset with a row of data for each day’s aggregated Apple Health tracking measurements.
What I’ve isolated is the additional amount of calories I’m burning on workout days compared to rest days. Noticeably, I’m not specifically isolating the calories burned during workouts. I’m getting at that in a roundabout, inexact way, so this is an area for improvement. A screenshot of the dashboard is below, but you can view the full interactive version of the dashboard here.
According to my data, I burn an additional 369.1 calories on days when I work out. This number is not perfect. On a few days, I have additional calories burned due to taking walks or mowing the lawn or whatever on top of working out that day. There are error bars, but it’s a decent estimate overall. Given that I’ve got data before and after macronutrient tracking, the trend is the more interesting part of analysis.
The trend is good. The seven-day rolling average shows that I’m burning more and more calories on workout days as time goes on. My hypothesis is that my additional protein intake is fueling better stamina during workouts, which would allow me to burn more calories as I work harder. The trend was going up before March 15th though, so it’s not clear that my hypothesis is correct.
Another way to go about comparing these pre- and post-March 15th data is a simple statistical test – a T Test. This test compares two datasets against each other (with some assumptions) and determines if the means of the two datasets are statistically different from each other. Using this test, the average additional calories burned on workout days pre- (341.4) vs. post (420.0)-March 15th are truly different from each other. This would support that my dietary changes are truly working, to the tune of about 23% more calories burned.
There are a few confounding factors – the type of class I took, the teacher teaching the class, and seasonality. As shown below the trend, there are definitely some classes and teachers that burn more calories than others. I will burn additional calories outside during warmer months by mowing the lawn or taking walks or whatever. I’m less likely during the winter to move more. Again, there are error bars with this analysis.
Two of the confounding factors can be explained. The type of class and teacher are not likely affecting these results. My workout schedule in terms of the classes I take, the teachers teaching them, and the days/times of those classes has been consistent for several months. However, it is likely that seasonality is affecting these results. There are a few spikes in May, like May 6th, in which walking long distances played a role in pushing up my additional calories burned on a workout day. Anecdotally, I doubt that seasonality is 100% responsible for the increase in additional calories burned on workout days. There are a couple spikes, but on many other days my workout class was the only movement / exercise I got.
You might also notice that the class types and teachers that burn the most calories are not the classes I attend the most frequently. If I wanted to burn even more calories, I should change my schedule to attend different classes. This gets into an important motivating factor of why I have continued to attend classes where I’ve failed before – I am choosing the classes that I like the most.
Two of the bottom three classes in additional calories burned, HH Crew and Row & Flow, are my favorite classes. HH Crew is rowing only, and Row & Flow is a mix of rowing and yoga. I love rowing, so naturally I like the class where it’s all rowing. I have limited flexibility and rough hamstrings, so Row & Flow is not only challenging but incredibly beneficial. R&F is on Fridays, and going through a yoga class on Friday evenings is a great way to end the work week. These other benefits are one reason why I have attended classes consistently, and that’s the most important thing right now. Even Row & Flow has a lot of strength training involved. It’s not easy.
Regardless, the early returns are encouraging. I have actual data to support my anecdotal descriptions, which is important for motivating me on my personal wellness journey. I feel stronger during workouts. I have more stamina. I’m recovering faster. I’m pulling faster times on the rower. I’ve pulled a few personal bests for 2,000 meters, the benchmark by which rowers measure their speed. I’ve rowed faster now than I did in college. I’m lifting heavier weights, albeit gradually.
All that remains is to keep going. Keep tracking. Keep exercising.