LL Andy Trapp, it’s great to be with you. You’re in our business school, affiliated with our data science program, and working on projects using science and technology to assist in real human need. You’re working on a really exciting project around refugee placement. Tell us about it.
AT Sure. This started several years ago. I started working with a fellow in the refugee space, Mike Mitchell, at HIAS—a refugee resettlement agency—and I started adding collaborators such as a professor at Oxford University, and we looked at this problem of how do you place refugees well? Presently, it’s a manual process, which is a whiteboard and people sitting in a room every week.
Operationally, they need to place the refugees that they receive through the United Nations High Commissioner for Refugees and the U.S. State Department. So these are already approved refugees. Where should they go throughout the country?
LL And you all were able to produce software that really optimizes that placement process, and makes it better and faster.
AT That’s right. It’s taken much of the labor and the manual decision making out of it, but not completely. And that’s an important yet subtle point.
LL To keep the humans in the loop.
AT You need the humans in the loop, because if you don’t have that, you can end up with recommendations that might not make sense. We can only approximately model the situation. But it’s really helped the decision maker to make better decisions.
LL I can imagine a lot of other kinds of applications for this approach, to optimize decision making in placement.You all are working on a lot of other things that are also associated with this way of doing things. Tell us about that.
AT Another project is matching our sophomore students to international project centers—we have roughly 50 project centers throughout the world.
LL A thousand students a year, now.
AT A thousand students that need matching. And we really need to understand … what do the students want? What are their preferences? Where would they like to go, or not go? So we have these different tiers: very interested, interested, and not interested.
LL And it used to be a manual process to place the students, and not very effective because we’d end up with so many students on wait lists. So this year was our first time deploying this new software.
AT Yes. Each student gets to recommend where they’d like to go. Also, project center directors have criteria they’re looking for, and so we’re matching in a way that honors the student’s preference and also according to their fit. Amazingly, we were able to achieve one hundred percent placement in terms of students going to “very interested in” project centers.
LL And this year you really did manage to do it and it was great.
AT It’s really a team effort.
LL So, speaking of team, you have students working on these various projects with you. Tell us about the students you find and how they work across these disciplinary boundaries.
AT It’s fascinating, being affiliated with the Data Science program at WPI, which includes statistics, business, and computer science. We really need these different skill sets and WPI is naturally aligned to understand technology in these different areas. So as we’re putting them together, I’ve found students with mathematical backgrounds or computer science or business—or all three. And it’s a joy to work with these students and plug them into problems. There’s no shortage of students that have a real passion to work on humanitarian and social challenges.
LL Well, that’s the great thing, right? We all think of it as being this super-high tech program in data science and industry out there gobbling up our graduates as fast as we can produce them. But there’s so many human problems that data science can be applied toward.
LL And it’s just great to see that work happening here at WPI. Professor Trapp, thank you so much.
AT Thank you.