Doughnut Run: Time Adjustments & Performance

Data Visualization

Data Wrangling

Joining / Merging

Modeling

Exploring race time adjustments and performance by cleaning, joining, and analyzing doughnut run datasets.

Authors

Affiliation

Tim Kashkarov

St. Lawrence University

Ivan Ramler

St. Lawrence University

Published

February 18, 2026

Module

Please note that these materials have not yet completed the required pedagogical and industry peer-reviews to become a published module on the SCORE Network. However, instructors are still welcome to use these materials if they are so inclined.

Introduction

Fun runs often include nontraditional race elements such as bonus challenges, costumes, or food stations. In a doughnut run, participants may stop during the race to eat doughnuts, and their official race time may be adjusted based on how many doughnuts they consume. This creates an interesting data analysis problem: how can we estimate doughnuts eaten from race timing data, and what relationship (if any) exists between doughnuts and performance?

In this worksheet, students analyze race data from a doughnut run using three datasets: adjusted race results, unadjusted race results, and a bonus threshold table. The main idea is to compare each runner’s adjusted and unadjusted times, compute the time difference, and use that difference as a proxy for how many doughnuts were eaten.

Throughout the activity, students will practice core data analysis skills including cleaning inconsistent time strings, converting time values into seconds, joining datasets by a common identifier, constructing a derived variable from thresholds, and exploring relationships with visualizations and a simple quadratic model.

This activity is especially useful for introducing students to realistic “messy data” workflows, where variables must be interpreted carefully and transformations must be justified before analysis begins.

Activity Length

This activity is suitable for a single class period or as a short out-of-class lab.

Learning Objectives

By the end of this activity, you will be able to:

Clean and standardize messy time values into a consistent format.
Convert time strings into numeric values (seconds) for analysis.
Join multiple datasets using a shared identifier.
Visualize relationships and compare patterns across groups.

Methods

Technology Requirement:

This activity requires R with familiarity in Quarto documents and tidyverse tools (especially dplyr, readr, and ggplot2). Students will also use lubridate to work with time data.

Data

You will work with three datasets:

doughnut2015.csv: adjusted race results
doughnut2015unadj.csv: unadjusted race results
doughnuttime.csv: the “bonus time thresholds” for each doughnut

Download Data:

[doughnut2015.csv] {target=“_blank”} [doughnut2015unadj.csv] {target=“_blank”} [doughnuttime.csv] {target=“_blank”}

The doughnut2015.csv dataset contains race results given the donut-adjustment. Each row represents one runner. Students will create time_sec during the cleaning process.

Variable Descriptions: doughnut2015.csv (Adjusted Results)

Variable	Description
Position	Overall finishing position
Race Number	Unique runner ID
Name	Runner name
Time	Race time
TimeAdj	Adjusted race tim
Category	Runner’s age category
Cat Pos	Position within category
Gender	Runner gender
Gen Pos	Position within gender

The doughnut2015unadj.csv dataset contains race results without donut-adjustment. Students will also create time_sec during the cleaning process.

Variable Descriptions: doughnut2015unadj.csv (Unadjusted Results)

Variable	Description
Position	Overall finishing position
Race Number	Unique runner ID
Name	Runner name
Time	Race time
Category	Runner’s age category
Cat Pos	Position within category
Gender	Runner gender
Gen Pos	Position within gender

The doughnuttime.csv dataset contains the number of doughnuts that determine threshold times.

Variable Descriptions: doughnuttime.csv

Variable	Description
Donut	Number of Donuts
Bonus	Time threshold

Data Sources

https://truetimeracing.com/event/doughnut-run-2015/ https://truetimeracing.com/event/doughnut-run-2015/ https://github.com/iramler/stat450-spr2026-score/blob/main/donuts/donut_times.jpg

Materials

Class handout
Solutions:
- Quarto with visible R code
- HTML with hideable R code

Conclusion

Through this module, students work with a realistic multi-table dataset and build a complete analysis workflow from raw data to interpretation. They begin by cleaning inconsistent time fields, converting times to seconds, and joining adjusted and unadjusted race results. They then use a threshold lookup table to explore how number of doughnuts eaten relates to race performance.

The activity highlights an important idea in applied data science: sometimes the variable of interest is not directly measured, and must be constructed from proxy information. This creates opportunities to discuss both the usefulness and the limitations of derived variables.

The final visualizations and quadratic model encourage students to move beyond simple linear thinking and consider whether relationships may be curved, noisy, or influenced by subgroups such as finish rank or gender. This makes the module a strong bridge between data wrangling and statistical modeling.

Additional Reading

Introduction to Data Wrangling with dplyr https://dplyr.tidyverse.org/
Data Visualization with ggplot2 https://ggplot2.tidyverse.org/
Working with Dates and Times in R https://lubridate.tidyverse.org/