It is now the start of my 7th semester (Fall 2022). Before getting in too far, I will take a moment to reflect on last semester (Spring 2022) and goals for the upcoming semester.
In the Spring 2022 semester, I taught ITCS 3162 (Introduction to Data Mining) once again. Except this time around, there were over 65 students as compared to my previous 12. In this upcoming semester, I now have around 76 students. I also had TAs for the first time last semester, which was amazing. However, this also introduced more things for me to learn like how to lead the TA team. Sometimes things were a little disorganized and chaotic, but I have hopefully learned from that so we can start the new semester strong. They have all done great so far and I’m very proud of their growth.
Course Structure Changes
The course structure remained mostly the same as written previously. This time, I added the post-class quizzes as per some of the previous students’ suggestions.
“The instructor is very understanding and courteous. They allow lots of flexibility in a good way. Since the course was project based, it did not feel like “busy work”, like many other classes do. I found the project based course to be more engaging and educational. There were no assessments but I believe that maybe some short quizzes would not hurt. Overall, it was one of my favorite courses that I have taken here at Charlotte.”
“I think some form of short quizzes (but not necessarily a large test) may help reinforce learning.”
– Some feedback from the Fall 2021 Course Evaluations
With this suggestion, a question bank was made for each module. The TAs were tasked with writing questions, which I then proofread and sometimes added to. The quiz then randomly selects five questions out of the question bank. The students get unlimited attempts in hopes that they retake it many times and eventually see all (or most) of the questions to reinforce their learning. It is also a good way for them to check their progress, and for me to see if there are any gaps in knowledge that may need some additional discussion. I believe writing the questions has also been good practice for the TAs in (1) assessing their own knowledge, (2) practicing communication/writing skills, and (3) understanding the students’ learning objectives for each module.
“I love the format of the class. The 5 somewhat open ended projects are very helpful with reinforcing modeling concepts and also increase my python coding skills. The quizzes were also a big help. The writing part makes you think more critically of what you are doing/have to do. The projects are also a big help to showcase your skills and what you have learned as you can directly put that in your resume.”
– Some feedback from the Spring 2022 Course Evaluations
For the lectures and exercises, we have also gone through various iterations of taking material from the previous semester and trying to make improvements (e.g., to the wording of material or adding new material). For the upcoming Fall 2022 semester, the course structure will remain the same.
My syllabi for the Spring 2022 and Fall 2022 semesters are below (both are near identical).
Responsible Data Science
In the Fall 2021 semester, we watched Coded Bias at the very end of the semester.
“Coded Bias says that there is a lack of legal structures for artificial intelligence, and that as a result, human rights are being violated. It says that some algorithms and artificial intelligence technologies discriminate by race and gender statuses in domains such as housing, career opportunities, healthcare, credit, education, and legalities.[5] Buolamwini and her colleagues were later asked to testify in front of the US Congress about artificial intelligence. Buolamwini subsequently created a digital advocacy group, the Algorithmic Justice League.[6]“
– Wikipedia on Coded Bias
The discussions were eye-opening and various students commented on their appreciation for showing this documentary. We are in a world where artificial intelligence is no longer a thing of fiction or where only very few people experience it. We live in a world where artificial intelligence is everywhere. It is used everywhere. And its insights are informing all aspects of society. As we are teaching new students about the algorithms, I think it is becoming increasingly important for them to also be taught how to critically think about their work, how it may be used, and any possible repercussions.
While this may not be the main focus of the class, I am interested in bringing in more related discussions throughout the semester, rather than just at the very end. While doing so, my goal is not to tell students “the answer,” but to encourage them to think critically and to give them confidence in becoming a part of these conversations.
Improvements to be made
Over the course of the previous semester, I took notes on things I wanted to improve upon. It was a very long list, and after talking with Dr. Saule about teaching, I’ve been told that this list of improvements never ends. It wouldn’t be realistic for me, as a continuing student, to try to follow through with everything sadly. However, I’ve prioritized this list for just a few changes that I’d still like to make:
- Adding more activities or discussions to help students get more practice with interpreting their results and telling a story about the insights.
- Adding the “impact section” to all projects with more class discussions leading up to give them practice with the critical thinking needed.
- Providing a list of example datasets for each project that students can use.
I felt like we spent a lot of time going through the algorithms, how they work, and the code to get them to work. However, we didn’t spend as much time in class discussing what to do after that or how to interpret the results. Students did this in their portfolio projects, but I think having some added class discussions would help better prepare them for the projects.
Second, we added an “impact section” in the Spring 2022 semester to their final project. This required students to think critically about their project and to discuss the possible benefits and harms that could occur. I felt like we did not prepare students enough for this section at the very end, and will now be adding it as a small piece to all of the portfolio projects. I will also add small class discussions to also give them more preparation. When starting these discussions with students, I really want to emphasize that we aren’t looking for right or wrong answers, but just to encourage them to think critically and to get practice communicating those thoughts.
Lastly, some students previously expressed anxiety about choosing a “good” dataset. While many others also enjoyed having the ability to choose their own data and questions, we want to start providing a list of possible datasets for each project. This way, students have examples to base their dataset search off of, or could possibly just pick from that list. The TAs are currently helping me put together a list of datasets for each project.
The above three are the improvements I plan to work on and have prioritized. Additionally, it may be nice to continue adding more questions to the quiz question banks (but a lesser priority). I would also like to review the lectures for linear regression and dimensionality reduction, as I worry if they were a bit shallow.
So far, I have greatly enjoyed teaching and am looking forward to the new semester.
Aileen Benedict 