AR Data Sciences
This Month’s Q & A: January 2022
Updated: Jan 19, 2022
Karl Schubert, Ph.D., FIET,
Professor of Practice and Associate Director of the Data Science Program for the College of Engineering, the Sam M. Walton College of Business, and the J. William Fulbright College of Arts & Sciences,
University of Arkansas
WHAT WOULD YOU do if it was your job—one of your jobs—to create a statewide IT/data science degree program so Arkansas students could begin anywhere and finish up anywhere, and all those credit hours at all those different colleges would count toward graduation, with none of them falling through the cracks? Oh, and one more thing: What if you received a portion of a $24-million-dollar five-year grant from the National Science Foundation and the State of Arkansas to make this happen…but you got the grant at the start of COVID-19, so you mostly couldn’t meet with all these statewide educators in person? Dr. Karl Schubert of U of A wrote part of the grant proposal for this project and spearheads the project, along with Dr. Stephen Addison of UCA, in conjunction with Jennifer Fowler of the Arkansas Economic Development Commission, which administers the grant. As we approach our second full year in pandemic mode, we asked Dr. Schubert to give us an update.
Tell me the idea behind this very ambitious initiative.
The initiative is called DART (for “Data Analytics that are Robust and Trusted”), and the idea for our education research theme was that we're too small a state not to work together. California, New York, Massachusetts—they would never consider doing this. But we’re Arkansas. We can do it. We can work together.
We can set aside the idea of competition for the benefit of the students, the state of Arkansas, and our employers, and agree that we're going to create, for example, a hub school in each quarter of the state. Not that it has to be only four hubs; we just need an institution in each part of the state so that students can start in a community or two-year college close to them and then finish in a four-year college close to them, if they so choose. By having these regions identified, we encourage the four-year hub schools to create concentrations and focused areas that are aligned with the companies in their part of the state. So, let’s say a student started and got the first two years somewhere in the Delta. “Well now,” the student might say, “I'm really interested in an agricultural concentration that's available up at A State.” Other students might just want to do a general concentration, like data analytics or data science, and they will be able to find that at probably any of the four-year schools throughout the state.
You make it sound easy—and I’m sure it hasn’t been.
One of the challenges we observed is that the different engineering programs in the state evolved over the decades, so they're all over the map on what they provide. Even if they're accredited, they're all over the map. So, what we decided—we being Jennifer Fowler at the AEDC and I, who wrote this part of the grant together—was that if we wanted to implement high quality data science programs across the state as quickly as possible, the best way to do it was to have a model. And since U of A was the furthest along in terms of developing a multidisciplinary Bachelor of Science degree with multiple concentrations, our program became that model. To be on this NSF grant in the education research theme, a school has to agree to more or less adopt, with only minor variations, the approach that we've set up.
How have the schools responded to that?
Normally, the schools all see themselves in competition with one another. So about three years ago, even before we’d received the grant, UCA’s Steve Addison, AEDC’s Jennifer Fowler, ACDS’ Bill Yoder, and I got together and talked about this competition problem. I described my hub-and-spoke vision for the state-wide program, and said that we would bring the two-year schools onboard by sharing everything we had with them.
We would commit to helping them put together programs whereby they could teach the first two years of the courses. I was committing us, and our deans were committing us, to help these schools by teaching their teachers how to teach these courses. The plan was that we would teach the teachers the first-year courses, and then those teachers we’d taught would teach the next set of teachers the first-year courses while we turned our attention to teaching the teachers how to teach the second-year courses—and so on, through the third- and fourth-year courses. And that’s what we’ve done. We’ve shared all of our videos for the courses, all our course materials, all our unit plans and lesson plans and lecture slides with the other schools—and then taught them how to use them. It's a first in the state, by the way, and NSF has told us that it is a first in their experience. As the two-year college folks have said, more than once, “Boy, we've never seen collaboration like this—ever.”
What part of this hasn’t gone as smoothly as you would like? Are you happy with the way the teachers are learning to teach?
I am. I think we would be further along with that, and they would be implementing faster, if we could do in-person meetings and teaching—even though we've been giving them opportunities to take other courses, to learn how to do programming, and so on. We met the first two times in person. We did the first session at UA Little Rock in September 2019, and then the second one at UCA in November 2019. Our third in-person session was going to be at Pulaski Tech, but then COVID hit and we went virtual. We did have an in-person teach-the-teachers workshop in Fayetteville last summer. I got special permission to hold it in person because we needed the interaction of the faculty. We needed them to be around one another. We needed them to be able to work on some stuff together to have the communication that I didn't feel we could we could do adequately in a virtual setting. We did report to NSF that there are a few things we haven’t gone as far as we wanted to with because of its being virtual.
So, I think we've been hampered in our rollout by that, although Laura Barry, who’s the person I work with at North Arkansas College, has already got students signed up. She's already “stealing”—that’s what she calls using all my course materials—and she’s watching the recordings from last year to see if she can pick up any pointers. I told her she's welcome to do that.
I think more schools would be coming along faster if things were different. I'm pleasantly surprised that A State put together a program modeled after ours really quickly. I mean, I went to see them right after the UA Little Rock session, and I spent an afternoon taking them through what we were doing. After that, they went dead silent. But then, all of a sudden, they had a program they’d put together to get a bachelor's degree going. It's a combination of data science and analytics, but it's modeled after our program. And it’s great to ASU’s credit.
On the other hand, I think SAU—South Arkansas University, in Magnolia—is much earlier in the process. We are working with them to focus on starting with the first two years of courses and building from there. We recruited them to be the fourth hub, joining us (the University of Arkansas), UCA, and A State as the hubs.
A while back, I heard that some of the schools were saying, “Oh, why do we have to do this, we can't afford this,” and so on. Have you gotten past that?
I think we have gotten past that. And the way we did it was to apply for a grant to get some of those schools some hardware. Steve Addison actually started working on the grant, and while it’s going to be a year or two before it comes to fruition, it will help immensely in defraying some of the costs of participating in this program. We’re also approaching industries in their regions to request assistance in funding for their labs.
We also decided to set up a virtual lab for those schools to be able to virtualize the instructional things that they needed, the student environments. Again, Steve has one of his lab people working on that and we hope we’ll be able to begin pilots for that this spring. And then, through the NSF DART grant, we've encouraged faculty members to apply for training support beyond the workshop that we’ve given them, because we didn't teach them how to teach the programming languages, we taught them how to teach the course that teaches the programming language, which is slightly different. They can apply for reimbursement for a class on R, or Python, or whatever. Jennifer and I have been reviewing and approving them for reimbursement to do that.
Also, with the help of Bill Yoder and ACDS, we're doing a statewide employer needs survey. We got approval for this from the Arkansas Division of Higher Education. In fact, they've changed their rules, so that if we do a statewide employer needs survey, which we’re going to do, it can be used by all the programs in the state to rationalize their proposed degree programs. Why should everybody have to do their own survey? We would all be surveying the same companies every time, so why not just do one survey and let everybody use it? I had my graduate research assistants create the survey, and then Jennifer and Steve and I finalized it and sent it to Bill for distribution to many of the businesses in the state.
We’ll also send it to companies that are regionally located. We can solicit funds from them to help us fund some of the hardware or lab purchases for smaller schools in their area of the state at the same time. That's the plan.
When do you expect this to be fully operational throughout the state?
We’re targeting to have full implementations by the end of the third year of the grant. That's what we're hoping for, so we can start the fourth year with actual students transferring from two-year schools to hub schools. We may have some students transferring in the third year, but the beginning of the fourth year—that’s the fall of 2024—is probably when it's really going to start happening.
Do you have a predictive analysis of how many grads you think this will add to the tech talent supply in the state?
No, I haven't done that yet. We haven't done a lot of advertising to this point. But I can tell you that in my first-year class last semester, I had 54 students. I didn't expect to have 50 students until at least next year. We had sized our program to be 50 students in a cohort and 200-225 in the program. And we’re probably going to have 80-100 students in the program when we start this coming fall. Going into spring of 2022, we have 82 students in the program and regularly have students (new and transfer) asking about the program.
So if I use that as an example and we start doing advertising and make it better known across the state that we've established a statewide program that provides our students this two-plus-two path to a data science career, I could imagine us graduating ultimately 300 or 400 students a year with associate degrees or bachelor's degrees across the state once we're at a steady pace. It's quite possible that that could be the case. When we get our four-year programs fully going and have the two-plus-two colleges up to speed, we could add quite a few.
Last, but not least, I should add that we are a fully “opt-in” program. That is, for two-year and four-year colleges and universities that want to participate, they can choose to join the program. The only requirement is that they follow the design of the state-wide program as we have defined it in the grant. It is really that simple and makes for a great collaboration for a great state.