Can Data Science help me run faster?
The rapid evolution of Artificial Intelligence (AI) has brought us on the verge of an era of unprecedented possibilities. Among the many breakthroughs, AI-driven technologies like ChatGPT have become incredibly popular because of their success in generating human-like text. One way that AI will help us innovate is in reducing the time it takes to gain insights from our data.
I decided to try this out with my own real world challenge, so I asked ChatGPT, “How can data science help me run faster?”
As of this writing, I am about 12-weeks away from running the New York City Marathon. In 2021, I ran the race in 3:28:17. In 2023, I hope to leverage my experience as a runner to beat that time. I also plan to leverage the power of data to train smarter and more effectively. Here’s how my training looked in 2021:
The charts above show my weekly mileage and training time (in minutes) for the 16-weeks leading up to the marathon. A few takeaways:
- I began the 16-week period running between 15–20 miles per week for about 3 hours of training time per week.
- My training program peaked about three weeks before the race at over 40 miles and ~6 hours of training time.
- My weekly training consistently increased over that time except for a 3-week period around late September where I missed some training time.
This quick and easy data analysis gives me a nice starting point for planning my training for this year. I have started off with a nice baseline over these past 4 weeks. Here is a comparison between 2021 in blue and 2023 (so far) in orange:
Each week, I plan to revisit the above data to monitor my progress. My assumption is that training further, longer, and more consistently will help me get faster. This raises another question — how do I know that I am getting faster?
The simple way to test this is to monitor my average pace over time and see if it’s improving. Here’s my average pace in 2021 for all training runs leading up to the marathon:
Interestingly, there was not much of a significant change in the average pace over the course of my training program — typically, it fluctuated between 9 and 10 minutes per mile. However, there were 2 notable exceptions in late September and early October where my pace was closer to my marathon pace (~8 mins per mile). Unsurprisingly, these were also from races (Bronx 10 Mile race in September and the Long Beach Island 18 Mile race in October).
My first takeaway from the pace analysis is that I should plan a few shorter races before the marathon to gauge my true race pace (as opposed to my training pace, which is slower and ranges). The second takeaway is that pace alone is not a great measure of performance in training runs as they vary in distance and intensity. Therefore, it will be important to identify additional metrics that can indicate performance, such as heart rate and stride cadence. I will dig deeper into those in my next post…
Until then, please consider donating to my fundraiser campaign for Friends of Firefighters. Friends of Firefighters is a 501(c)3 not-for-profit organization that provides independent, confidential, and free mental health counseling and wellness services to active and retired FDNY firefighters and their family members. Each year, participants in the NYC Marathon raise millions of dollars for various causes like Friends of Firefighters.
