NUS Module Reviews 2023/2024 Business Analytics (BZA), School of Computing (SoC) - Y2S1: BT2101, CS2040

I took 7 courses which are 22 units in total during this semester. Among them, three are core courses (BT2102, CS2040, ST2334) for my major, two are under the GE pillars (GEX1014, GESS1025), the remaining two CFG mods are under the unrestricted electives and I won't discuss them in this post.

I had also enrolled in IS4250 but dropped it after week one since I could not cope well with so many courses.

BT2101 Econometrics Modeling for Business Analytics

Introduction:

Link to Nusmods: https://nusmods.com/courses/BT2101/econometrics-modeling-for-business-analytics

Semester 1

Lecturer: Dr. Varun KARAMSHETTY (VK)

Timetable Workload (lecture + tutorial + lab): 2 + 1 + 0

Attendance: Lectures have participation, which are tracked through PollEv during lectures.  It is based on submission instead of accuracy. Tutorials do not have attendance. So attend the lectures and you can earn 5% on your grade.

Lecture Recordings: Videos are made available on the following Monday.

Grading Scheme / Assessment:

Lectures:

Lectures are held in campus. Since there are in-lecture PollEv questions, it's advisable to attend every lecture.

Tutorials:

Tutorials start in week 3. TAs do a brief revision and go through the solutions of the individual PS or group PS submitted on the previous week.

Individual PS:

According to my lecture slides, there were originally supposed to be 10 individual problem sets (PS) in total, but I only had four on my laptop, so I'm not sure. I think Prof VK might have made some changes and canceled the individual PS for the second half semester. 

One PS was released each week starting from week 2. Tutorials were used to discuss the solutions to these PS. Each PS included one to four questions, with each question broken down into a few subquestions. Students usually need to run code on given datasets using R and analyse the results, although some PS did not involve coding.

The individual PS were graded based on submission rather than accuracy. This means you'd receive full marks as long as you submitted something that showed effort. Overall, the problem sets were relatively easy and provided another free 5% on your grade.

The group PS will begin in the second half of the semester, and the details are in the group work session below. In addition to the individual and group PS, there are optional PS which are not part of the syllabus of this course and do not contribute to your grade. One optional PS is released each week starting from week 2 throughout the semester. You will only receive answers if you submit responses to these optional PS. I completed 5 optional PS myself and found them to be useful to learn beyond the syllabus. I recommend trying them only if you have time or are interested.

Exam Settings: 

Midterm was conducted in week 7, 2 hrs, held in school campus face-to-face via Examsoft.

Endterm was conduced in week 13, 2 hrs, held in school campus face-to-face via Examsoft.

Exams are open book with no internet connection. Students can bring as much printed/handwritten material as they can carry and soft copies are also allowed. There are MCQ + True/False + Descriptive questions but I'm unsure about the components of each question type. 

I actually didn't do well on the midterm because I wasn't familiar enough with the contents and struggled to use my lecture and tutorial notes effectively. For example, when I read through a question, I often recognized the theories behind but wasn't confident about the answer. I felt like the professor or TA had covered it at some point, but I couldn't remember which lecture or tutorial it was from. 

Therefore to prepare for the end-term exam, I created a "dictionary" for all the content covered during the semester. It's essentially a table of contents that allows me to quickly locate the relevant notes and remarks. For example, if a question asks about panel data, I can search for "panel data" in my dictionary and see that it was covered in the lecture notes for weeks 8, 9, and 10. Then, based on the specific question, I can identify the relevant material and review my notes accordingly. It helped me efficiently find and use the information I needed and I eventually did well for the final exam. (The final exam scores were not disclosed to students. But based on my overall grade and the scores I received for other components, I deduced that my final exam score was quite good.)

Of course, you don't need a dictionary like this if you can memorise and understand everything, and I'm sure there are more effective ways to achieve this. I just wanted to share this method in hope that it might be helpful to those who need it.

Group Work: 

There are a total of 6 group problem sets (PS). Grading for the group PS is based on a combination of your group-level performance (graded on the accuracy of your submissions) and your peer-review scores (graded on the quality of your individual contributions). 

Each week, one PS will be released. Every two weeks, students will be randomly assigned to a group, and together, they will complete one PS per week. Thus, each student will have the opportunity to work with three different randomly assigned groups. Professor VK explained that this helps prevent students from being disadvantaged by group members who are not contributing and encourages collaboration with a diverse range of students. 

Additionally, after each two-week period, students are required to complete a survey to provide feedback and assign scores (0 - 100) to their group members. The grading team will use these surveys to assess the peer review components. Rest assured that the scores given by your group members will serve only as a reference during the assessment of the peer review components. It does not mean you need perfect scores of 100 from all group members to receive full marks. Based on my friends' experiences and my own, as long as the overall scores are good, you will receive full marks for the peer-review scores.

Personally, I don't see much difference between the individual PS and the group PS. The formats are essentially the same (involves coding in R and analysis). The only distinction is that the content in the second half of the semester is more complex compared to the first half. Thus, There are more questions in the group PS and they are harder. Working in groups helps reduce the chances of making mistakes on the PS, as your group members may catch errors you might overlook.

There are many ways to allocate tasks within a group. For example, some groups might divide the work so that one student handles questions a-c, while another tackles questions d-e, and so on. However, I highly recommend that all group members attempt the entire PS. Since the questions are interconnected, it makes more sense for everyone to work through everything and discuss the final answers together before submission.

Conclusion:

The workload is manageable, and the theories and analyses covered are intuitive and easy to grasp. The coding required is not difficult, and Professor VK usually provides the necessary codes, which is a relief if you're not particularly good at coding like me. Professor VK is very approachable and humorous, and he made a strong effort to keep the lectures engaging and enjoyable. Additionally, all three tutors for my semester are very kind and eager to assist.

CS2040 Data Structures and Algorithms

Introduction:

Link to Nusmods: https://nusmods.com/courses/CS2040/data-structures-and-algorithms

Semester 1

Lecturer: Dr Chong Ket Fah

Timetable Workload (lecture + tutorial + lab): 3 + 1 + 2

There are two lectures per week, one is 2 hrs long and the other one is 1 hr long. Lectures are online through Zoom, Tutorials and Labs are onsite.

Attendance: Lectures do not have participation. While tutorials and labs have attendance.

Lecture Recordings: Videos are uploaded after the live lectures if I remember correctly.

Grading Scheme / Assessment:

Lectures:

Well, I attended (or watched recordings of) every lecture and still screwed up this course. Apparently actively learning is more important than just passively listening. So what if you just can't learn effectively from the lectures? Good question, I'd say each individual has their own best way of studying. If you're confident in your methods, stick with what works for you. However, if you're struggling and unsure of how to improve, attending the lectures or watching recordings can still be helpful, even if it’s just placebo.

Here is a youtube video showing one of the algorithm covered for this course (Stack and Heap). You may find other relevant videos in the same playlist. They may be helpful to better understand the algorithms if you want to do some self-studies beforehand.

Tutorials:

Tutorials start in week 3. Besides the attendance, students need to answer or raise questions during class to earn the participation grades. Tutorial questions tend to be more theoretical, focussing on the problem solving skills rather on hands-on skills comparing to the labs exercises. They are generally formatted like exam questions.

Labs:

Labs start in week 2. Students will be divided in groups to work on discuss the in-lab assignment (ILA). Each week, one ILA will be released per week and there are 10 ILA in total. 

In addition to the ILAs, there are 8 take-home assignments (THA), with approximately two THAs released biweekly. During lab sessions, TAs will provide a brief summary of the lecture content, discuss THAs, and explain the THA for that particular session. Students are then required to work with their group mates on the week's ILA and submit a copy of pseudo-code as a group. This submission will be used to mark lab attendance.

ILA will be released at 10 am on the day of the lab session and the deadline is set to 10 am on the next day. So if you can't pass all the test cases for you ILA, do consult your group mates or TAs ASAP. Kattis auto-grading is used to calculate the marks for lab assignments.

 Programming style stands for:

1. Modularity
2. Meaningful comments
3. Meaningful/descriptive identifiers
4. Proper indentation

There will be 0.5 mark deducted if programming style is terrible on all of 4 main categories. This means if you pass all the test cases, you should not have marks deducted unless your coding style is really terrible.

In general, labs focus more on students' programming skills (ability to translate idea/algorithm into actual program).

Online quizzes:

There are two online quizzes held on VisuAlgo and conducted during lab sessions on week 7 and week 13 respectively. Each quiz takes about 30 minutes and is open-book. 

You can always practice on VisuAlgo as many times as you want if you're not confident about the quiz. However, keep in mind that VisuAlgo has a fixed set of question types and only randomly assigns values to the questions. To make it easier, you may consider record all the question types and their respective solutions while you exercise. During the quiz, you just need to substitute the given numbers into these question types and calculate the results.

Exam Settings:

Midterm (week 7, 1.5 hrs)

Final (week 13, 2 hrs)

Both midterm and final were open-book, conducted on campus, face to face using Examplify. The exams focus more on students' problem-solving skills, eg. ability to understand and reason about the problem, ability to apply knowledge to formulate solution.

There are MCQ + Analysis + Structured questions but I'm unsure about the components of each question type. Analysis questions will give you a statement and ask you it's true or false and the reasonings. Structured questions ask for pseudo-codes. 

Given my own experience with the course, the best advice I can offer is to manage your time well. Make sure you don't end up with so little time that you can't even read the last few structured questions.

Conclusion:

If you're aware that coding, algorithms, or anything related to CS2040 is not your strong suit, it might be beneficial to do some self-study before the term starts. 

Moreover, if you encounter any issues with the lab assignments, make sure to consult your TAs before the deadline to resolve them and aim to score full marks for the lab components. The same advice applies to quizzes. Since the midterm and final exams can be quite challenging, they are actually the main determiner of your grade. It's crucial to ensure that your performance in other components does not pull you down.