🏃 Ultra-Marathon Race Analysis – Exploring Gender, Age & Seasonal Trends

For this project, I analyzed a massive real-world dataset of ultra-marathon races as part of a self-directed learning goal to improve my data analysis and storytelling skills using Python. The dataset, sourced from Kaggle, contains over 7 million race records spanning two centuries, offering rich insights into human endurance and athletic performance.

To keep the analysis focused and manageable, I worked with a subset of the data — specifically 50K and 50Mi races held in the USA during 2020.

🔍 Project Goals

My main goal was to turn messy, real-world data into clear insights through thoughtful cleaning, analysis, and visualization. Along the way, I challenged myself with the following questions:

• Do male and female athletes perform differently in ultra races?

• Which age groups tend to perform best — or worst — in 50-mile events?

• Does the time of year (season) impact how fast people run these races?

🧼 Cleaning & Preparing the Data

The raw data required substantial cleaning. I:

• Filtered for U.S. races in 2020 at distances of exactly 50 kilometers or 50 miles

• Cleaned up inconsistent text data (e.g., race names, performance times)

• Removed irrelevant or redundant columns

• Converted time and speed values into usable numeric formats

• Engineered new features like season, derived from race dates

This gave me a clean dataset with well-structured fields like:

• Athlete age

• Gender

• Average speed (km/h)

• Race length (50K vs 50Mi)

• Season (Winter, Spring, Summer, Fall)

📊 Visual Storytelling

Using Seaborn and Matplotlib, I explored the relationships between gender, age, and seasonal performance.

📍 Key Insights:

• Speed by Gender:
  Male runners had slightly higher average speeds in both 50K and 50Mi races, but the difference was not as large as commonly assumed.

• Age Group Performance:
  The most competitive age groups for 50Mi races fell between 30–40 years old, with older and much younger runners showing lower average speeds. I limited the rankings to age groups with a minimum number of participants to ensure fairness.

• Seasonal Patterns:
  Races held in summer months saw a slight drop in average speed, supporting the idea that extreme heat may impact endurance performance.

Each of these insights was visualized using tools like:

• Histograms and KDE plots

• Violin plots split by gender

• Regression lines showing age-speed trends

• Bar charts ranking age group performance

📦 Tools Used

• Python (Jupyter Notebook)

• Pandas & NumPy for data wrangling

• Seaborn & Matplotlib for visualization

• Kaggle for the original dataset

🧠 What I Learned

This project helped me deepen my understanding of:

• Working with large, real-world datasets

• Cleaning and transforming data for meaningful analysis

• Communicating complex ideas visually

• Deriving actionable insights from filtered, structured subsets

It also reinforced how much storytelling matters in data — not just what the data says, but how you present it can make all the difference.

👀 Explore the Work

If you're interested in the code and methodology behind this analysis, you can check out the full notebook on my GitHub.