Creating these pipelines—and learning how to harness their power for good—is the core of IS 455.
Inside the Classroom
Australia’s famed pipeline waves summon surfers from around the world, but it was a church mission call that drew Mark Keith to the land down under. Australia introduced Keith to many wonders, including Microsoft Excel, software he first encountered while assisting his mission president. Keith realized that he wanted to study computers, but he never imagined that someday he would teach students about the unseen pipelines that power today’s technology.
Now a BYU Marriott associate professor of information systems (IS), Keith teaches IS 455: Predictive Data Analytics, which focuses on machine learning pipelines. These pipelines are composed of steps that data passes through to be analyzed, ultimately leading to a final prediction or outcome. Pipelines drive product recommendations on Amazon and movie suggestions that pop up on Netflix. Creating these pipelines—and learning how to harness their power for good—is the core of IS 455.
Keith stumbled into machine learning during his PhD program at Arizona State University. His dissertation led him to analyze the same set of data again and again, looking for a specific outcome. “I could never get that dataset to tell a story, so I had to start over. But because of that process, I got really good at knowing how to take a big set of variables and find out what the true cause-and-effect relationships were,” he says. “I’m great at slicing data every different way.”
This skill led him to create the first machine learning class at West Texas A&M University and then at BYU Marriott, where he started teaching in 2012. The content is now spread over two courses—IS 415 and IS 455.
IS 415: Machine Learning—a course Keith developed that is currently taught by assistant professor Katy Reese—covers topics such as the programming language Python, basic statistics, and techniques for data storytelling. “I couldn’t teach IS 455 without Katy Reese laying the groundwork in IS 415,” Keith emphasizes.
The second class, IS 455, fittingly resembles a pipeline itself as Keith guides students through a series of steps. First, the students scrape and pull data off the web to predict numbers and categories. Then they put this data together into a pipeline that makes recommendations. The most recent class worked with data on Legos to produce predictions about which Lego sets people would want to purchase.
Rachel Yorke, a student from Lakeville, Minnesota, who took IS 455 in winter 2024, appreciated the diverse applicability of the skills she learned in class. “As Professor Keith talked about the different ways that machine learning could be applied, it made me think about how I could use it day-to-day to make life better,” she says.
The course showed Yorke how machine learning can change lives, not just purchasing decisions. “One of the most impactful examples we talked about was Professor Keith’s work with another college that was struggling with high dropout,” she recalls. “He created a pipeline to look for indicators and trends among students likely to drop out.” These predictions enabled professors and staff to proactively reach out and support those students.
IS 455 not only helps students master pipeline development but also helps them become responsible stewards of this powerful tool. “Like all technology, machine learning is divinely inspired,” Keith says. “This technology is going to help people learn faster. As our students use it sincerely and honestly, they are going to do more than they’ve ever done before.”
______
Written by Shannon Keeley