My Favorite Student Projects of 2025

In a project-based data science class, I’ve gotten student work that has delighted me, confused me, horrified me, and sparked genuine curiosity in me. Here are my favorite Data Science project ideas from the 2025 school year.

Project #1 - Dunkin Donuts

A few months into the school year, my student discovered this article from the Boston Globe (sorry about the paywall). In essence, a hobbyist and student at Tufts named Jack Burton created a quick lookup tool for various Dunkin items to compare the price of that item across several locations (or find the location where that item is cheapest). My student, herself an admitted Dunkin addict, wanted to expand on this idea and create maps of the variation in Iced Coffee prices around the country.

Mr. Burton, helpfully, provided his source code on his github page, so we found out how to use the Dunkin API to scrape prices from Dunkin location data, and my student took it from there to create maps, visualizations, and descriptions of Dunkin pricing around the country, around Boston, etc.

I loved this project from conception to follow-through. First, it was clearly possible, which can be a challenge for students. Many times, students want to do the impossible with their data projects: create a perfect bracket, solve a systemic issue of injustice, predict the next Oscar winners - all terrific goals that inevitably lead to disappointment when students discover the deep complexity of the data (or lack thereof). Second, this project was ambitious in non-obvious ways. We had never covered working with location data. My student needed to learn new skills that built on the ones covered in class. And third, this project was engaging. I’ve been criticized in the past for allowing students to pursue projects that feel “silly.” I would posit that this student will gain more skills by pursuing a project she really wanted to do, rather than dive into a rubric and check boxes. She will have many opportunities to use those same skills in ways that have larger impacts, and this exploration of Dunkin prices may have been her doorway into that reality.

Project #2 - Meat and GDP

A particularly ambitious student of mine wanted to explore the relationship between meat consumption and economic factors. This student - sort of a goofball and class clown type (but an overall terrific and sweet kid) - seemed to think the topic was really funny. And then, he started to explore the data.

Using the CIA World Factbook and some data on meat consumption from Kaggle, he started to explore trends in meat consumption by region, individual country, and overall. He tracked meat consumption’s parity with GDP and encountered plenty of data conundrums along the way. How do you visualize multiple line graphs at once with clarity? How do you group data in different ways? Of different data types?

This is a real snowball situation. What may have started as a good goof turned into a legit project, one that sparked curiosity and meaningful learning. It’s my favorite kind of learning, and it’s my favorite way to interact with data - with some whimsy and an open mind.

Project #3 - Movie List

Have you ever met a spreadsheet kid? These kids are rare, and I consider them to be kindred spirits. At some point in their lives, someone showed them a few Excel tricks. Since then, they keep spreadsheets on everything. One of my students this year is a spreadsheet kid, and she has kept a running spreadsheet of movies she has watched for years. The spreadsheet contained hundreds of movies, including watch date and her own personal rating out of 10. She wanted to do something with this spreadsheet, and so I recommended merging it with IMDB data that would provide genres, runtimes, director, actors, awards, average ratings etc. From there, some natural questions arose: What “awful” movie did she love most? What “amazing” movie did she hate most? Do her genre preferences change depending on the year? Have they changed overall? Is she more inclined to rate movies with female directors highly? And many more.

I describe projects like this as Data Journals - taking an aspect of one’s life and building data models on top of it to understand it better. There’s a lot of personal growth and reflection baked into projects like this, and it allows for a uniquely metacognitive experience. To access that personal reflection, she needed a fair amount of data science - merging, working with datatypes, statistical functions, visualization, etc. But some of the best data science projects are the ones for which the data science is merely the vehicle to some broader or deeper understanding. I love the idea that a data project can be surprising or inspiring.