Q&A on MADS capstone course with lecturer Elle O’Brien
Just like many of the Master of Applied Data Science (MADS) students she teaches, lecturer and research investigator Elle O’Brien transitioned to data science after beginning a career in another field altogether. For O’Brien, who grew up in New York, that field was computational neuroscience. Her initial attraction to data science in graduate school, she said, had a lot to do with salary.
“I was happy doing science research, but post-docs in biology-adjacent fields didn’t pay very well, and I wanted to pay off my big student loans ASAP,” O’Brien says. “Data science has a lot of action, it’s fast developing, and it has lots of worthwhile problems to solve. I started pivoting that way midway through my PhD, just getting a sense that I could be happy in data science.”
After learning how to use machine learning libraries in personal projects and independent studies, O’Brien used these methods in her dissertation as she could and sought out other opportunities to go beyond.
“In hindsight, it was a huge time commitment to do this and a PhD, so I see a program like MADS potentially saving people some time,” she said.
After spending time after graduation making software tools to help data scientists, O’Brien gave a special presentation to a SIADS 643: Machine Learning Pipelines class, and — driven to teach and impressed with the MADS program — she joined the MADS faculty in March 2021.
This term, O’Brien is teaching SIADS 697: Capstone, a project-based course in which students propose and build end-to-end data science projects in their domains of interest. Students are asked to demonstrate mastery of data science concepts and methods from their MADS training and produce a creative, original and technically rigorous portfolio piece. Projects are supervised by instructors throughout the course with regular peer review.
Here, O’Brien provides some insight into course process, projects and what sets the MADS capstone apart.
Q: What can MADS students anticipate from the capstone course process?
O’Brien: “By the time MADS students get to the capstone, they’ve covered a lot of topics in data science and tend to have ambitious ideas for how they can 'make their mark' on the field (or in their domain, be it finance or climate science or social science). The capstone is like a big workshop or sandbox where they can team up and pursue those dream projects. The teaching team is more like coaches or consultants at this point — students can book time with us to get our opinions and advice, but their project is very much on their own terms. Some teams are building dashboards, and others are working with scientists at the university to bring a data science perspective to a research question, and others are more interested in the storytelling and visualization angle.
“Students team up and send a proposal to the teaching team before the capstone officially starts, and then they have all eight weeks to work on it. The result is a blog post or research paper that’s ready to share with the world. It’s a chance to make a portfolio project that can help you land a job, launch your personal brand online, or bring some value to an organization or cause you care about.”
Q: What are you most looking forward to about teaching capstone this term?
O’Brien: “Seeing what people do. MADS students are an amazing bunch, and a lot of folks had jobs and careers before coming to data science. Those interests blend with data science skills to make projects that are original and inspired.”
Q: Have students shared any capstone project ideas with you that you’re excited to see?
O’Brien: “So, so many. A few students are extending methods from Sabermetrics, a data scientific approach to baseball, to financial markets. Another group is analyzing tweets from international embassies to understand foreign relations. Another is working with the ACLU to help them find trends in their outreach program. Yet another is working on building machine learning models for medical decision making that are interpretable to human experts. I could really go on!”
Q: What sets the MADS capstone apart from other programs’ project-based courses?
O’Brien: “We’re trying to strike the perfect balance between enough guidance that things get done, and they get done well, with the freedom to pursue something unexpected, original and close to students’ professional goals. And we have chances for students to get input from teachers with experience in data science, data engineering and software engineering almost every day of the (work) week! Another difference is that we’re really aiming for students to make something outward facing. The goal isn’t just to wow your teachers, it’s to make something that moves your career ahead and contributes to the broader world of data science.
“Everybody will be hustling in the last two weeks to make their portfolio pieces impeccable, and they’re sure to wish they had even more time — but in the end, having something done and ready to ship is worth more than perfection. We’re going to make sure students have that and the experience of taking a data science project start-to-finish. I hope people see their capstone project as a useful milestone on their resume five and ten years down the road.”