I am a 53 year-old sculler who was bitten by the rowing “bug” 35 years ago while a freshman at Boston University, where I
primarily coxed 8s and sculled. I participated in freshman varsity and, although I had water in my veins from growing up in a small town on the Northshore of Long Island, where I boated, sailed, swam, and fished with my father, rowing as a sport permeated to my very core.
As a healthcare informatician and analytics officer in a medical device company, real-time data measurements are a key part of my life and substance. I run a blog at http://johnrzaleski.com that covers aspects of medical analytics and, also, my experiences with rowing & rowing analytics. There are a number of excellent sites on the web that provide knowledge, software tools, and guidance on usage of data (for example, http://rowsandall.com).
How to automatically collect rowing data
In this article, I am going to take a deeper dive into some of the data that can be collected automatically using some of the products out in the market and how to interpret what is collected. There is a nice comparison blog entry on Rowperfect written by Editor-in-Chief Rebecca Caroe, dated May 22nd, 2017 (“GPS comparison: NK, Coxmate, Minimax”) that provides a functional comparison among the various GPS technologies. Key aspects of these technologies is the ability to collect data for download, principally for post-workout analysis. Performance measures such as clock speed update rate provide indications as to how frequently data can be collected from these devices. The details of these aspects are covered well in the aforementioned piece.
Yet, once the data are collected, what are we looking for, and what value is in all of these individual numeric data points?
In order to provide some insight on this front, I’m going to take the reader down the road of my personal rowing partner, the Garmin Vivoactive HR and its data set (photo).
I row in vicinity of the Northern Chesapeake Bay, on the Elk River, which is one of the five major rivers that feeds the Chesapeake (these include the Susquehanna, Northeast, Bohemia, Sassafras, & Elk). My workouts typically range from 6,000 – 9,000 meters, depending on circumstances. When unable to row outside, I also will row on my Concept 2 PM5 unit, which sees most use in the winter months. During these workouts I normally track the following measures:
- Heart rate
- Speed or pace
- Stokes per minute
- Distance per stroke
- Stoke power
- Overall time
Through the GPS, I also track the actual route taken. A summary of a subset of these data are shown in the following figure, which depicts the actual track of the boat, the pace, heart rate, stroke rate and distance per stroke. These data show the typical workout (casual) for the overall rowing session of 1 hour 8 minutes and 4 seconds (elapsed). From these data, the average heart rate is derived (58 beats/minute) with a maximum of 109 beats/minute. Total strokes was counted at 1400 with an average stroke rate of 21 spm and a max of 34 spm. Average distance per stroke was 5.68 meters with best pace at 2:28 minutes per 500 meters. Total distance covered was 8,070 meters.
The story behind rowing data
The data tell a story over time, and the key is that having the data lets one analyze the entire workout and assess points of weakness, fatigue, strength, as well as adherence to form.
In my case, I can tell when I am in proper form as I see the pace open up and the distance per stroke lengthen. I govern my stroke rate in order to ensure proper form over speed, and to ensure that I don’t rush the slide up to the catch. Also, particularly as I normally row alone, this lets me monitor proper form of the oars when I drop them in at the catch and ensure proper extraction at the release.
In the plots below, you can see points where values go to zero (for instance, stroke rate). This occurs at points of inflection where I am turning around (for instance) or avoiding an obstacle. One thing we have on the Elk are many motor boaters and kayakers, and depending on time of day, one can meet up with many. The region is tidal so my workouts need to be timed with wind and tide.
Comparing and combining data points
The data also allow for correlations among variables. For instance, determining whether there is an association between stroke rate and distance per stroke. Furthermore, understanding conditioning is an important part of the analysis: the relationship between heart rate and stroke rate (for instance), or heart rate and temperature or time of day. These are all factors in determining performance. This particular rowing session was conducted early in the morning (starting approximately 6 a.m.)
A benefit of linking the rower data with the GPS data is the ability to tie in wind, temperature, and tide to provide a more holistic picture of the experience, which is more realistic when considering actual rowers out on the water.
Other aspects of the analysis include calculating power
This is done automatically on an indoor rower, such as the Concept2. Yet, on the boat this requires a bit more physics. I aim to address this aspect in a future blog post. Oxford University’s Anu Dudhia discusses the basic physics of rowing, and describes the mathematics associated with estimating the power and energy required to move a shell through the water, which depends on drag of the boat and other factors. In my next post, I intend to take this on and make more presentable to the layman-rower as well as discussing the specifics of how to download and access data for analysis.
John R. Zaleski, Ph.D., CAP, CPHIMS