Can data science meaningfully change your life? Asked differently, can you use easily available sensors and apps along with common data and analytics tools to collect enough data about your daily life to craft a sustainable action plan? The answer is yes. Here’s how.
I’ve been on a personal journey for about five years, and it has taken me from euphoric heights to the depths of depression and back again (several times). It’s a journey many people know well. Every morning I’d look in the mirror and decide that today was the day I was going to lose weight. But more than that. Armed with the knowledge that diets are bulls#$t, I was going to change everything about my lifestyle: no more bread, no more pasta, no more alcohol, no more desserts, no more refined sugar … today was the day I was going to make it happen. The steely resolve empowered by my silent, personal pep talk usually lasted until lunchtime (although occasionally I’d be “good” until dinner). It will come as no surprise that I have struggled with my weight for the better part of my adult life.
The Quantified Self Movement
As I wrote in a recent article about Apple Health, “From May through October 2013 I put myself on a 1,500-calorie-per-day diet and logged every morsel of food I put in my mouth using MyFitnessPal. I committed to purposeful walking 4 miles every morning (one hour per day), using Runkeeper and Spotify to keep my pace under 15 minutes per mile, and I lost 55 lbs. This was the most remarkable accomplishment of my life (to that point). But my euphoria was short-lived.
By June of 2017 I had put back 45 lbs. of the 55 lbs. I had lost. This time I was truly helpless. Every medical professional I went to see told me the same thing. ‘You’re fighting gravity. Your body wants to be a certain weight, you’ve trained it to live on 1,500 calories per day, and if you want to lose weight, you’ll have to learn to live on 1,200 calories per day or less. And you’ll need to step up your exercise.’ To make matters worse, the doctors I was seeing did not want me to try to exist on 1,200 calories per day because that is considered a starvation diet and is known to be unsustainable.
I was told over and over again that all I could do was what I was doing and just hope for the best.”
At the time, Fitbit, Jawbone, Misfit, Basis, etc. were all the rage. They all counted steps and attempted to monitor sleep. From a practical perspective, the data they generated was all but useless. You could walk 20k steps one day and 2k the next and 13k the next. These data points made great anecdotal fodder, but you had no way to understand what was causal, what was corollary, or what action to take next. That was then.
Today, the sensors and apps available are exponentially superior to the sensors and apps that started the Quantified Self Movement, and with the advent of Machine Learning, you can do more with them.
In May 2018, I sought some medical advice. What would it really take to conquer this? More importantly, assuming I could lose the amount of weight I wanted to lose, could I ever possibly keep it off?
My Side Hustle
Today is the 72nd day I have been within +/-5 lbs. of the weight my body wants to be. This is an exceptional milestone for me and it truly has me thinking about changing day jobs. As side hustles go, this one has been amazing. I’ve used all the tools at my disposal. Data-driven marketing (DMM) tools, analytics tools, data science tools and methods, nutritional information from the government and private vendors, and machine learning (AI) for pattern matching (in the data) and image recognition (to take pictures of food and automatically log it – this is harder to do than it sounds, and it may be a while before it’s truly useful).
I’m working on an app that automatically logs what I eat and collects data from my Apple Watch, MyFitnessPal, Runkeeper, Ski Tracks, and my digital scale. It turns this data into actionable insights and can inform specific choices about food consumption and exercise requirements in real time. It is highly predictive and very helpful in understanding what is likely to happen (from a health and wellness perspective) on any given day.
Here’s a screenshot of my dashboard from this morning. So far, there are 28 data points that are used for each day’s calculations. Using this data, I can predict what I will weigh tomorrow within 25 basis points of accuracy. I’ve used the information in this foundational set of metrics to create specific action plans for each day – it takes the guesswork out of the lifestyle choice I’ve made. (Notice I did not use the word “diet” – remember, diets are bulls#$t!)
These numbers by themselves don’t tell the whole story. They only show progress. The dashboard uses moving averages to smooth the data and monitor trends rather than daily data points. Trends matter because daily weight fluctuations can be caused by any number of factors unrelated to the number of calories you have consumed.
You’ve heard of the USDA’s “food pyramid.” It was taught to every American in grade school. It’s complete nonsense.
We all need a certain amount of protein, carbohydrates, and fat every day. It doesn’t matter if you’re into Keto or Paleo or Atkins or Diet Center or WW or Jenny Craig or Gluten Free or Vegan or anything else, for that matter. We all need to figure out our unique requirements. This is a huge part of my side hustle. Eating appropriately sized amounts of fresh, healthy food that correspond to the macro nutrient balance I personally need has been a key component to my success. It may work for you too.
Am I Really Giving Up My Day Job?
I’m told that if I keep the weight off for two years, I’m likely to keep it off permanently. But two years isn’t a mental picture I can use, so I’m treating this as a “one day at a time” lifestyle choice. Today, the data says do ___. So far, so good.
If this keeps working, I will share more of my journey, recipes, algorithms, data collection tools and methods, and ultimately my planning tools with everyone. If this turns out to be really valuable, I can’t think of anything that would be more rewarding than gifting this to people who might benefit from it. I’ll keep you posted.
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it.