This post is about 2 startups, a PhD, a long distance relationship and using Heart Rate Variability (HRV) to understand life stress.

This post is about looking back at the past 15 months of my life and trying to figure out how to make things better.

This post is about being an entrepreneur, a toolmaker and a data scientist.

This post is about being human.

The Story

Not that hard if you move to a town where the temperature is always the same.

15 months ago I left and sold almost everything I had, and packed the remaining 70 liters of stuff in a backpack, before taking off to San Francisco.

With a group of friends, we started Bloom Technologies, a digital health company, focussed on helping expecting mothers having a healthy pregnancy.

Like any other startup, it’s been a rollercoaster and an amazing adventure. Something unique and hard to describe.

The air you breathe in San Francisco.

At the same time I’ve been also busy working on finalizing my PhD on applied machine learning at Eindhoven University of Technology, and running another business, HRV4Training, helping athletes optimizing performance.

Working hours piled up fast.



A night out at the pub. Piling up 300+/hours per month, peaking at 16/day (RescueTime).

For lucky, passionate people, these are the best times.

Working on a Friday night in a bar is perfectly normal. Working on a Sunday morning in a garage is the least you can do. We all love what we do. We are all dreamers.

Work is almost everything.

Personal life, anyone?

For more than a year, this was my relationship with Alessandra, my wife:

You know it’s a startup when you have post-its behind you all the time. If you are in a long distance relationship, get the Without app. It’s awesome.

Not helpful.

My living situation was pretty much like in Silicon Valley, living and working at our startup house (I did definitely have the best BBQs of my life — thanks Eric).

Back to Europe?

7 beds in 12 days are no fun for your back.

Going back to Europe was no better, since I had no home anymore, and I was sleeping at each one of my friend’s places across Europe (thanks everyone). Fun, but stressful.

Life stress piled up, until the inevitable happened.

My productivity crashed:

I almost went down to the infamous 40 hours/week (or ~160/month).

What did my physiology have to say?

Here is where the data comes in. As a toolmaker and data scientist I’ve been working for a long time on wearable technology to monitor and make sense of physiological data.

What about my breakdown?

Can we capture life stress in today’s frenetic world and possibly make better decisions in the future?

The Data

I packaged most of what I learnt in years of work on wearable technology and data mining in simple tools (apps) that can accurately measure Heart Rate Variability (HRV) using just the phone’s camera, and help understanding our physiology better.

Camera HRV at work.

Long story short, HRV is regulated by the autonomic nervous system, and when measured under the proper conditions, can reveal a great deal of information regarding how our body reacts to stressors (social, physical, psychological, environmental, etc.). You can find more information on the various aspects of HRV on the HRV4Training blog.

Personally, I’ve been mainly using HRV data to guide day to day training, since acute changes in HRV reflect quite well the intensity of our aerobic workouts. However, this time I wanted to look at something different.

I wanted to look at HRV baseline with respect to my ability to cope with life stressors. For this I needed first to quantify my wellbeing. For someone that spends all his time in front of a computer, RescueTime (a software logging all you do on your computer) was the answer.

Just looking at weekly summaries data I could see all I’ve been up to.

1 week low productivity? Holiday.

2 weeks low productivity? Ill.

5 weeks low productivity? Total breakdown.

RescueTime was my perfect indicator of wellbeing:

When you live in front of your computer, you can figure out pretty much everything by looking at RescueTime data.

HRV & (Physiological) Stress

Making a few oversimplifications, we expect HRV to be higher when we are in a better physiological state, less stressed, ready to perform. On the other hand, we expect HRV to be lower when we are more stressed.

Looking at my HRV (Baseline rMSSD in the plot), I could clearly see it was crashed during weeks I was not productive (threshold at 4 hours/day of high productivity).

On the other hand, working many hours/day was typically associated with higher HRV.

I found this relation very interesting, since in “stress research”, often more working hours are expected to be associated with lower HRV, since more work means more stress (and stress is often thought of as negative).

Relation between weekly baseline HRV (computed using HRV4Training) and weekly highly productive hours as logged by RescueTime over a year.

In my data, the relation was inverted.

There was a really strong correlation between my baseline HRV (7 days moving average) and more highly productive hours, as logged by RescueTime.

In today’s society things can be very different. I was working many hours trying to build something I believe in, and that was definitely positive stress. I am not implying any causal relation here, but the fact that a better physiological state (higher HRV) was always associated with my ability to work effectively for much longer, makes a lot of sense.

This relation holds when controlled for flying/jet-lag, sick days, baseline heart rate, fitness level, training load and training performance (all factors that can also influence HRV).

What does this mean?

During the past 15 months, life was very intense, with many periods of hard work and high productivity and a “burnout”.

Baseline HRV was highly correlated with my productivity, meaning that my physiology reflected very well what was going on with my life stress and my ability to work effectively. The relations between these variables are less clear (what comes first?), and will be part of my future analysis. I am interested in, for example, trying to determine if the influence of life stress on my HRV can be picked up before there is a drop in productivity, and possibly use an early HRV drop as a sign that we should try to change something.

Looking at this data was eye opening. I always expected physical stress (exercise) to have a stronger effect on our physiology (or HRV) with respect to psychological stress (or what I called here life stress). This belief was mainly due to the fact that any sort of physical effort completely disrupts our physiological values (higher HR, lower HRV, etc.). However, when looking at baseline HRV measurements averaged over a week, for a person that is not a professional athlete, values were much more representative of life stress and of the overall load I was putting on myself.

Written by

Founder, Data Science @ouraring Lecturer @VUamsterdam. PhD in Machine Learning, 2x MSc: Sport Science, Computer Science Engineering. Runner

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store