Recently, I had the privilege of attending the North Carolina Data Science Education Summit. One of my goals there was to discuss the role of mathematical modeling in data science education. During my time developing and running data science programs at Ramapo College, I frequently heard from employers about how valuable these skills are for their data scientists.  I would reframe that to say that they are looking for data scientists who have a mathematical modeling mindset.  

At the conference, I reflected on an important question: How can we help teachers cultivate this mathematical modeling mindset and guide students to integrate it into their data science activities?

A Real Example to Ground the Mindset

To illustrate, I drew on two COMAP mathematical contest problems: the 2023 MidMCM problem and the data from 2025 MCM Problem C. Both focus on analyzing Olympic medal counts. The adapted version is given as: Adapted 2025 MCM Problem C Data

The next Olympic Games are being hosted by Los Angeles, CA from July 14 to 30, 2028. The International Olympic Committee (IOC) must order enough medals and flags for the medal ceremonies. The IOC must make sure they have enough medals and flags for all the ceremonies to avoid being embarrassed. The IOC also does not want to waste money, however, by ordering too many medals and flags that will not be used. Therefore, the IOC has asked your modeling team to provide models they can use to determine how many medals and flags they will need.

How Modeling Courses Approach the Problem

For a mathematical modeling course, the next steps might be to ask questions like:

• What you need to solve the problem?
• What questions do you have?

There is likely data that they need to solve the problem and this way they have a chance to think through what type of data they may want and how they may use it.  It allows students to develop a strategy and think about potential pathways towards a solution.  They must think critically about what information they need, what assumptions to make, and how to structure their solution. This process encourages deep problem formulation and scaffolding, which is the essence of a mathematical modeling mindset.  Note that many of the approaches will involve using data.

How Data Science Courses Typically Approach It

For a data science course, we could ask the same question and provide the data set. 

In my own courses, students would open the data in Python and begin data exploration.  They may create charts, graphs, and start predicting on different variables.  The hope is that these efforts will help them generate a solution.  However, often they find other interesting information, but it may or may not address the initial question.  So, before having students plow into the data, it is important that they understand how and why they are using it. 

In my own classes, I’ve noticed something interesting: students who start without data often end up identifying the same types of data provided to the other group, but only after engaging deeply with the problem first. This initial thinking step makes their problem-solving more effective and intentional.

Bringing the Modeling Mindset Into the Classroom

So, I challenge both teachers and students: How can we bring this mathematical modeling mindset into the classroom? One simple strategy is to have students formulate and think through the problem before they begin wrangling and wrestling with data. While this may not always mirror real-world scenarios, that thinking process goes a long way toward building stronger problem-solving skills.