Wade Fagen-Ulmschneider is a Teaching Associate Professor of Computer Science at The University of Illinois at Urbana-Champaign (UIUC). With a passion for data, he has taught a variety of data-centric courses (Data Structures, Data Driven Discovery, among others) and works with students on numerous data visualizations that have accumulated over 25,000,000 interactions. He has been selected as one of the National Academy of Engineering's Frontiers of Engineering Education scholars; awarded the Collins Award for Innovative Teaching; and he has been consistently ranked as an excellent instructor by his students for the past ten years.
Karle Flanagan is a statistics instructor at the University of Illinois at Urbana-Champaign. She has taught Statistics 100 to thousands of students at UIUC since Spring of 2014. She also serves as the MS advisor for the statistics department. In 2018, she was awarded the Illinois Student Government’s Teaching Excellence Award and in February of 2019, she also won the Campus Award for Excellence in Undergraduate Teaching. She completed her undergraduate degree in mathematics, with a minor in secondary education. After earning a Master’s in Statistics at UIUC, she started teaching Stat 100 and is currently working on course development for other advanced courses and data science courses using Python.
Wade & Karle have provided their expertise in data science to several GET summer school programs, including the General Education program from Shanghai Jiao Tong University in Shanghai, China and the UNIST program from Ulsan National Institute of Science and Technology in Ulsan, South Korea. Wade & Karle share their experiences co-teaching innovative data science classes at Illinois.
When the best of both worlds collide
KF: Wade is known on-campus for creating viral visualizations. And when he was working on one of his more recent visualizations, he kind of discovered me through that.
WFU: We actually were able to collect a bunch of data about what's the best course at Illinois. I looked at that list that we were able to put together, and there was a bunch of people that I recognized. And then, there was this statistics professor who I had never heard of.
WFU: We actually ended up meeting up for coffee to discuss how we taught in the classroom - like, “Why do students really like my class? Why do students like your class? How can we share ideas to make the best of both worlds?”
KF: A couple of deans and then my department head asked me to go to a conference at Berkeley, and one thing I learned was that they do a lot of work with statistics and computer science, so they have those professors definitely work together. So, I told Illinois, and they said, “Do you know anyone in computer science that would be interested [in working] on this?” And I said, “I know the perfect person!”
On co-teaching courses at Illinois
KF: We really like having both of us teach in every single lecture. We get a lot of feedback from the students that they like seeing the interaction between us. And, it makes it more interesting to have an expert in computer science talk about computer science, and have an expert in statistics talk about statistics. We try to make the lectures as interactive between the two of us as possible - and with the students.
WFU: We don’t want this to be something that you can just get off of YouTube or social media. If you’re in our classroom, we want you to be fully engaged. We want, if you’re coming here, to have an amazing experience that really is creative, that’s doing, that’s active. Just having that bouncing off of each other, having the energy of each other, kind of questioning and puzzling and solving problems together is a lot of fun.
On their course, Data Science Discovery
KF: It’s a new course at Illinois, and it’s the first semester that it’s being offered, as a quantitative [general education] requirement. We developed it last year because Illinois really wanted to have a gen ed data science class. We kind of take the approach with Data Science Discovery that we want any student that comes to Illinois to be able to take the class. The students don’t have to have any statistics background or programming background. We designed it so that anyone can do well in it.
WFU: We have students in political science and art theatre and psychology - something like 70 different majors are in Data Science Discovery this semester here on-campus. And it’s just so fantastic to see how, being able to manipulate data, being able to look at it, being able to visualize it in interesting ways is super powerful to so many different disciplines.
On their GET summer school program experience
KF: We piloted “Data Science Discovery” last semester with 20 students from 20 different majors - to make sure that we were doing everything right. People literally from anywhere on campus could take this course and be successful. Now, [fall semester 2019], it grew by a factor of 10. We have 200 students.
WFU: We had a lot of great feedback, and we were able to partner with GET to have a version of this class to teach to students coming here to Illinois for a couple of weeks. We took the best bits of our 16-week course and compressed it into a 2-week program that focuses on: “How can you be creative? How can you use data? How can you do amazing things with data?
On storytelling through data visualization
KF: One thing we really focused on, in the GET course, and in our course this semester, is using data to tell a story. We said, “Find a data set, and figure out something interesting about it, using everything you’ve learned in our class.” They picked data sets they were interested in, they did the requirements for the project, and went above and beyond. They seemed to have fun with exploring their own data and telling stories about whatever they found.
WFU: One of my favorite presentations that I saw from them looked at the perceptions of how words differ across cultures. We call this the perception of probability words. So, if I say, “We believe this is true,” or “It is very likely that it is true,” different people in different cultures actually think of those phrases as having different meanings.
WFU: Some people believe that “we believe” is actually a very weak statement because beliefs are weak and they are not supporting facts. But other cultures feel that “we believe” is actually a very strong statement; that it’s kind of a personal conviction you’re really standing behind. So, just seeing that difference and seeing how, as we survey different people, [we can] present data in a way that really shows off how cultural differences can happen. Words have such strong meanings when you convey probabilistic statements.
WFU: It was fun to see the students take that on and see how different cultures come together, how communication and visualization of this information is such an important tool, especially globally.
KF: We teach our course with Python, and [the GET students] used Python to do a bunch of different things - even things beyond what we taught them. Some of them had programming background in other languages, and they incorporated them, as well. We didn’t want to give them too many rules, and it ended up going really well.
On quantifying uncertainty
WFU: The way we solve problems is different, so we often try to bring both approaches in. I have an engineering background. I went through an engineering course; I think like an engineer. But Karle has a background in statistics; she’s a mathematician. It’s a completely different field. And she brings that knowledge into class.
KF: There’s some specific problems, in general, that we like to use that we both solved differently. When we were prepping for the lecture, we realized, “Oh! We are doing this problem completely different, but both answers are correct. And, so I think the students like seeing that because it gives them an interesting perspective. They can see that you can actually solve the same problem in two different ways.
WFU: One thing that Karle has always said is that, “Statistics is making decisions under uncertainty.” All throughout our life - we don’t know what’s going to happen throughout our day. There’s a lot of uncertainty in life. And data science is figuring out a way to quantify that uncertainty and enabling us to make the best decisions under uncertainty.
Thanks, Wade & Karle!