Some random notes about approaching your master thesis if you are a student of MSc Data Science at University of Potsdam. In no particular order. (I put together this list for some juniors and friends who kept asking me questions about the thesis)
-
Create a draft version of your thesis from some latex template (This HPI one is recommended) as soon as possible and start writing the Section names, Sub-section names and initial bullets points for the main content. It doesn’t need to be perfect. This ensures you have some place to note down your thoughts in a structured way. Later you can modify/improve upon or remove these draft sections.
-
Define your Research Questions (RQs) based on the scope and work you are planning to do. I recommend max 2-3 RQs. Get those reviewed from your supervisor. These RQs are the most important initial step as you have to write about your approach and results basically answering these research questions.
-
Try finding some master thesis online which is close to your topic or is in similar domain. Read it entirely. This would give you some ideas on how other people approached these kinds of problems and went on writing about it etc. basically it will give you some structure for your own thesis.
-
If you have to come up with a new approach/algorithm in your thesis, later you will have to compare the results of this algorithm with some existing state-of-the-art approaches. These are generally termed as Baselines. Discuss which baselines to consider for your thesis with your supervisor. Do extensive literature review in order to make sure you don’t skip some obvious state-of-the-art.
-
It is recommended to keep the scope of your thesis smaller and diving deep in it rather than having a larger scope and not able to explore all dimensions of the scope.
-
Make sure you send your thesis draft for review to your supervisor at least 2 months before the submission deadline, so they can give you some feedback.
-
You will have to write a background chapter (generally chapter 2 of your thesis) introducing all important topics related to your thesis. For example, if you use Natural Language Processing in your thesis, you have to explain 101 of NLP. I recommend watching lots of youtube videos for these and taking notes in your own language. Later just use these notes to write polished version of your Background chapter. I highly recommend youtranscription.com to transcribe all the videos you are watching and keep track of the transcripts.
-
For Plagiarism check and Grammar suggestions, I recommend that you buy 1 month grammerly premium subscription at the later stage of your writing phase.
-
Drink lots and lots of clubmate. :)
Hope this helps! All the best!