1st semester finished. Tips for future students. Master in Data Science, University of Warsaw

This story is full of tips probably useful for those who are going to study a Master program in Data Science at the University of Warsaw.

10 min readMar 15, 2020

A little bit of history

Since 2018 I was looking for a graduate program in Data Science in different European universities. I will write later about the selection process.

If you also tried to search that, you will probably agree that not so much is written about the Data Science master programs. Simply, they are very new. Considering the fact that I had a job in Warsaw and the University of Warsaw had evening classes I thought it is the best option to study here. Well, I do not regret it :)

It is August 2019… now I am finalizing organizational details to start studying from September.

But…

What about 1st semester? What about the subjects? Do I need any prep work? Which professor to choose? Python or R? What to take into consideration? How are the exams? Homeworks? Projects? And many what and how and why or when…

Well, you can not find the answers to many of them before attending the first class. Yeah…

On the one hand, it is understandable as the program is very new and coordinators have a lot to do.

The only information source for me was a girl from Latin America who I found on the internet. She was already in the 2nd year of the same program. She was very kind to answer the tones of my questions. May God bless her for the help 🙌! As of my return to that girl’s help, I feel I should write some of the tips that could be useful for those who will start studying in the 1st semester.

However, all writing here is my own thoughts and not official opinions. For some of you, tips may be useful but for some of you not.

Subjects of the 1st Semester

Generally, the first semester was very hard. But it depends on your background. If you have just graduated from economics or computer science you may find it easy. The program is combining subjects from computer science, statistics, and economics.

The full list of the subjects of my 1st semester is this:

R: intro/data cleaning and imputation R/basics of visualization
Statistics and Exploratory Data Analysis
Python and SQL: intro/SQL platforms
Unsupervised Learning
Introduction to Data Science
Applied Microeconomics
Applied Macroeconomics
Communication and Autopresentation
Intellectual Property Protection
Workplace health and safety

For each class I will write “difficulty level” based on a class average grade.

1. R: intro/data cleaning and imputation R / basics of visualization

Difficulty level 8 out of 10. Considering you are a beginner in R. Structure of the program is here. (Link may already be expired). You may start learning R at DataCamp or somewhere else to make your life easier.

I think the teacher is running on the program at the same speed as that car in the photo:

So you need to:

And not only. You may need to do exercises at the end of each class material. Become a friend with DataCamp. It will help you.

The exam is very hard because of the time constraint. 90 minutes and 4 big tasks. You may need to run very fast on the first task which is a list of ~16 small questions and spend maybe 30 minutes on that. The other tasks have also quite big weights and if you solve them you will get a whole 10 or 16 points. Students tend to use up to 1 hour on the first task to answer all 16 questions and get 1 point for each answer rather than leaving very time-consuming questions and going to other big tasks 2, 3 and 4.

There will be an exam mock-up before the real exam so you may need to solve that to prepare well. The teacher will give you feedback on the mock-up, it was very useful.

The exam is open book. You can use notes. That helped me a lot! I hope it will help you too. Especially notes from mock-up exam ;).

The whole grade depends on your exam.

2. Statistics and Exploratory Data Analysis

Difficulty level 9 out of 10. Grade’s 20% is from the presentation of any paper about statistics and 80% from the exam. The exam is an open book.

When I wrote the exam I thought this is the best exam I have ever written this semester.

But…

Almost 50% of the class failed. Thank God I got 3. What a disappointment. The criteria for the exam’s evaluation is not clear. :( Should I write Shakespeare's poem? It is statistics, you got the answer or not. What I heard is that you need to elaborate on answers as much as you can. Try to write at least two lines of the answers. I wish I knew that before.

Yeah, the presentation… Make sure you get the right paper, get the professor’s acceptance and seriously — DO NOT speak more than 10 minutes. My grade was deducted by 3 points mostly because we spoke more than 10 minutes. So good luck! And do not spend a lot of time to prepare a presentation. I know a group who spent 2 days in preparation and got 16 out of 20 and a group who spent 2 hours and got 18.

3. Python and SQL: intro / SQL platforms

Difficulty level 9 out of 10. Grade’s 60% comes from the project and 40% from the exam. The exam is a closed book.

Project… The best suggestion is to find a committed mate with whom you will create a web app based on Python. I was lucky. I had a great mate. If the teacher suggests using Bottle for a web framework, please do not use it, instead use Flask or Django. The Bottle does not have good documentation and tutorials are very few.

Requirements for the project are that you must use SQL to have at least 5 tables interconnected, web framework (Bottle or other) and Python as a back-end, have a login and logout functions. If you show some stats in the app derived from your SQL database you may get the higher grade. You will probably be asked to write a 10+ pages report and make a presentation at the last day of the course.

Better to get in touch with your friends who know Python or web development. They may be the ultimate helpers to solve some problems in your app. Oh, and God bless StackOverflow. And also, Erim’s (my mate) cousin is a great guy, he helped us and taught us many things.

Some inspiration: I love this Flask tutorial

Try to be in contact with the teacher and let him/her know what you are doing in the project. You do not want to found out at the presentation that your project is crap. Mine was… so we needed additional work. But it worth. We learned more :)

For the exam, make sure you know how to write queries in SQL, INNER joints, simple SELECT queries. Also, you should know how to write functions in Python and class objects, inheritance. What else? I don’t remember.

4. Unsupervised Learning

Difficulty level 7 out of 10. My professor was awesome. The grade consists of 25% for presentation and 3 papers. Each paper grade weighted max 25%. During the course, we covered 3 topics. Clustering, Dimensionality Reduction and Association Rules. For each topic classes last 5 weeks. At the end of each topic, we were given 2 weeks to submit a paper. A paper should have 8+pages where you use any dataset and apply the algorithms we learned during class. I even published a story based on one of our papers on Towards Data Science blog.

Choose a mate who is committed to work hard with you and write a good paper. Find a dataset on the internet: Kaggle.com or https://datasetsearch.research.google.com/.

The presentation will be fine, you should choose an interesting topic.

5. Introduction to Data Science

Difficulty level 8 out of 10. Be cautious about the brutal exam. 12 deadly questions will decide your grade.

I am sorry to discourage you but it is better to know this in advance. Professor will have a presentation based on the slides he/she will share. It would be better if I learned slides by heart. Seriously. I do not have many tips to tell you here.

6. Applied Microeconomics

Difficulty level 7 out of 10. I always favored Macroeconomics until I attend the class in Microeconomics conducted by Professor Tomasz Kopczewski. He is the light at the faculty. His passion and dedication will awake your curiosity about science.

His goal is not to fail you (unlike other professors) but to teach you! You will for sure love his experiments during the lab and lecture.

He will teach you a lot from behavioral economics and will let you use many Data Science algorithms in Microeconomics. Try not to miss any class. His English is not good but if you bear that you will enjoy the class. Seriously.

You can choose the weights in your grading system. The default split is 25% assignment/paper, 25% activities on the forum (professor will create a special forum) and 50% exam. The exam is open book. I suggest you put 100% on the assignment and work on it well. Choose the right people in the group, talk with the professor and get his acceptance on the topic. I was not lucky, professor hated our topic. But we missed agreeing on this beforehand. Prepare a draft and let the professor give you feedback. Use anything that the professor taught you during this course which will increase the chance for you to get a better grade. If you work well on the paper, there is a big chance to get 5 in the overall course. Also, please note that there is a mandatory minimum to get 60 points together in assignment and activity and a minimum of 30 points in the exam. This may change in your syllabus.

If you do not have a background in Economics you may need to read Hal Varian’s book about Economics. Also, here is a video I enjoyed watching about basic concepts in Microeconomics.

7. Applied Macroeconomics

Difficulty level 10 out of 10. 30% of the grade comes from Homeworks and 70% from the exam. The (deadly) exam is a closed book.

Above I mentioned that I loved Macroeconomics. But…

However, there is always some positive side. The best part of the course is that you will have labs in Python! Woohoo. Yeah, Macroeconomics modeling in Python.

So what you can do? Try to solve all the Homeworks. Those students who solved Homeworks I hardly remember anyone to fail. By the stats, every year there are 20% or more people failed. Materials are very tough. Do not expect you will understand more than 50% of materials during the class. You will probably need a lot of work here.

My tips are to create a study group out of those students who understand Economics (or want to work hard to understand) and work together. To prepare for the exam, split the problems in the group. Solve and share within your study group. Then discuss and learn together. That helped me a lot. Sometimes people prefer to do it individually. It’s up to you.

I was also lucky to find the genius in the class (my classmate from India). She helped me a lot to understand homework problems and also some of the exam problems.

8. Communication and Autopresentation

Difficulty level 1 out of 10. This class is a lot of fun. I think because of the professor. She has very interesting tools that you will find very funny. But… I think some of the tools and methods that you learn during the class will be very useful for you. You will also practice speech at the class many times and get some feedback from the professor.

The brand of the professor’s laptop is Ferrari. I don’t know how it happens. Maybe you can ask.

You need to read two books and watch five movies to pass the exam. Just kidding. You do not need to worry about the exam. I do not remember anyone who got less than 5 in this subject. Show up at the exam and write anything the professor asks. Probably to describe someone's personality and go home happy. You got 5.

9. Intellectual Property Protection

Difficulty level 3 out of 10. Read the slides professor shares on his website. Do not miss the exam and pass 13 questions out of 26 (If I remember correctly). Not difficult but without reading some laws and rules from the slides you may find the exam hard.

10. Workplace health and safety

Difficulty level 2 out of 10. For this course, there is a specific online platform that you should go there, find the course and register. Monitor the deadline for passing the exam. You do not need unexpected retake :) You should read slides to pass 30 minutes exam on the platform.