A graph that describes my velocity for three heart rate levels over time. Interestingly my velocity at low heart rates (red line) has been increasing quite gradually, while my velocity at higher heart rates has had a serious dip in the second half of 2016 (when I became a dad) and started increasing again in the second quarter of 2017.

Data Science Cyling — Part 2, fitness over time

In my previous post I did a visual analysis of the relationship between heart rate and velocity. It always is a good start to get a feeling for the data when starting any data science project. My initial analysis already pointed out that the relationship between velocity and heart rate goes a bit wonky if we do not take the slope of the terrain into account.

The original goal was to get an idea of my fitness and how it develops over time. In this post I will do this in two ways. I will do a basic aggregation as a descriptive analysis to see how a couple of fit indicators change over time. Secondly I will do some basic modeling on the relationship between heart rate and velocity and investigate how this model changes over time.

Investigating my peak performances over time

So one of the most basic indicators of fitness is your peak performance, or the maximum speed and heart rate that you can keep for a set amount of time. Most training programs revolve around Functional Treshold Heart Rate (FTHR) or Functional Treshold Power (FTP), which are the maximum heart rate or power you can sustain for 60 minutes. So similarly I thought it’d make sense to look at a couple of indicators similar to this idea.

Because I have no power information, I am bound to look at heart rate and velocity. To do…