Strides and Statistics: A Deep Dive in Strava Data Insights
As some of you may know, I am an avid runner and enjoy working on data projects. I extracted my running data from Strava to put together an analysis spanning the past three years. My running journey began in 2019 and I started tracking my performance in 2020 during the onset pandemic.
There are two predominant philosophies around endurance training (1) training for distance or (2) training for time. Opting for time-centric goals enables a runner to have a degree of flexibility and adjust intensity as needed. On the other hand, distance training can assist runners with working on pacing and tracking progress from a mileage perspective.
In my personal training regime, I oscillate between the philosophies and train on intuition — if I begin my run and feel that speed comes naturally that day, then time is the focus. If speed doesn’t feel natural but I have endurance, I pivot to covering distance.
In 2020, the bulk of my training consisted of solo runs, given stay-at-home restrictions during the pandemic. I maintained an average pace of 9’48” per mile and ran 3,195 miles for the year. A deeper dive into my training data highlights an uptrend in my speed on a yearly basis: an improvement of 10.4% in 2021 compared to 2020, a further enhancement of 1.4% in 2022 relative to 2021, and a 2.3% improvement in 2023 vs 2022. The drivers for this change in pace are due to practice over a long time horizon, introducing strength training and running with groups.
From the school of thought on time-focused training, I have a box-and-whisker plot that delineates the distribution of my running time on a quarterly basis from 2020–2023 year-to-date. I had the most pronounced dispersion in running time during the second quarter of 2020. The following quarters show a trend of diminishing variability, with running times clustering around the median. My outliers on the upper bound were largely from group training or preparation for an upcoming race, on the downside, the outliers were from rest weeks or periods of time I was injured.
Analyzing my training program from a distance / mileage standpoint, below is a linear regression illustrating the interplay between mileage and speed. In Strava, ‘max speed’ is calculated by the peak speed recorded between any two GPS coordinates during run activity. The trend that I wanted to highlight here is that there has been an improvement in my average ‘max speed’ on a year-over-year basis. From 2020 through 2023 my arithmetic mean ‘max speed’ was 6.05, 6.12, 9.72 and 10.89 respectively. I believe there was a positive relationship with training with clubs and strength training.
Fitness has always been a large part of my life. I stumbled into this habit as a teenager and have been an enthusiast ever since, participating in wrestling, weightlifting, climbing, and running. Three years ago, I began my running journey, and it has been a life changing experience. I am thankful for the friends, mentors, and community I’ve gained throughout the process.
