Do data scientists need a Masters

Which master is right for me? (Data Science / Machine Learning)

Which master is right for me? (Data Science / Machine Learning)
Hi forum,

I am currently studying mathematics in the fourth semester of my bachelor's degree at the University of Bonn.
I like the applied areas of statistics + numerics and I'd like to do something in the field of data science / machine learning on my job. Many skills, especially programming routines, should be acquired while studying. However, it would be an advantage if my master’s degree covers a lot of relevant knowledge / skills.

My options (which I've thought of so far):

1 master's degree in mathematics in Bonn:
  • + prestigious university
  • + interesting modules in discrete mathematics
  • - Hardly any specialization in statistics is possible
  • -Master modules have little connection with data science
2 Masters in Mathematics elsewhere
  • + Various options in statistics / numerics (e.g. in Aachen / Heidelberg)
  • -You need pure math modules, which are unlikely to be relevant
3 Master's in Scientific Computing Heidelberg, module handbook (my previous favorite)
  • + Here, too, a lot of freedom of choice in statistics / numerics
  • + Data Science Module
4 Master's degree in Statistics TU Dortmund
  • + in-depth knowledge of statistics is not useless in the area of ​​data science
  • -little freedom of choice
5 Master's in data science TU Dortmund
  • + Name
  • -relatively new course (I think)
  • -little freedom of choice
Which master is best? Is the subject area of ​​the Masters relevant for later jobs at all? Is there a demand for data scientists in Germany?

I look forward to answers



Edited 2 times. Last on 06/30/17 6:04 pm.
Re: which master is right for me? (Area of ​​data science / machine learning)
As you mentioned earlier, data science is just coding. No master will help you. Especially not math. Find a company that is looking for people like that and working as a student trainee, and pray that you don't just have to do SQL. The most important thing in this area is work experience.
There is definitely a need, but most of the applicants do not meet the expectations.

In the US, for example, every Hinz and Kunz recruits data scientists, but every manager tells you that they can't find any. Too high demands.
Already know physicists (Americans) who have still not got a data science position after 400 applications.

In Germany, people tend to do LogRegression or are more of an engineer position, i.e. Hadoop or something (although hardly anyone uses that) If you're unlucky, you still do SAS or SQL all day, nothing with machine learning. The problem is, data science has to be promoted by supervisors to be effective, but most of them have no idea about it. And a lot of money has to be invested, which many are not ready for.

The best example is Amazon. Anyone who knows a little about marketing strategies knows that Amazon dwarfs the competition. They solve all their problems with machine learning.
In e-commerce, you can no longer survive without data science. Banks are replacing more and more business and economists with data scientists.



Edited 1 time. Last on 06/30/17 7:07 pm.
Anonymous  📅 30.06.2017 19:58:00
Re: which master is right for me? (Area of ​​data science / machine learning)
@imperius
I wouldn't completely sign the first part like that, but we've had the topic before. Data science doesn't work without code, but there are also some that only SAS and R can do. But that differs from place to place, whereas without the statistics behind it, you don't get very far at all. However, practical experience is definitely important.

Here is an example:
Hardly any programming: http://jobs.heise.de/Job/Data-Scientist-m-w.95718085.html
almost only programming: http://jobs.heise.de/Job/Data-Scientist-m-w.78283505.html

I also agree to the second part.

@TE
I would take that in Heidelberg, freedom of choice and beautiful Heidelberg make the difference.
Re: which master is right for me? (Data Science / Machine Learning)
Let's take a look at the first job:
  • Processing of large amounts of data, data mining and model building as well as creation of system prediction and system optimization models
It's just programming.
  • Visualization of data analyzes and derivation of recommendations for action for departments
Programming, then meetings.
  • Providing real-time analytics to decision makers
I have no idea what they mean by that, I also do real-time analyzes, it's just programming.
  • Execution of maturity level analyzes for BI as well as BI business value analyzes
No idea, but I always get a shock when I read BI, I would avoid such passages.
  • Methodical guidance in the field of analytics, especially for exploratory and predictive analyzes
Programming only

The second job isn't even data science, I wouldn't apply for it. More like software development.

I agree, however, that you have to have a clue about statistics, but even 90% of what you learn in a statistics course is not needed. It also depends on your area of ​​application. Therefore it doesn't really matter where the TE is studying. Internships / working students are more important.
You can tell, however, that the companies have no idea about data science. Most job advertisements just suck. Those from American companies usually make more sense.



Edited 1 time. Last on 6/30/17 8:09 PM.
Anonymous  📅 30.06.2017 21:22:22
Re: which master is right for me? (Data Science / Machine Learning)
  • Processing of large amounts of data, data mining and model building as well as creation of system prediction and system optimization models
Of course there is programming involved, but that doesn't mean that the data scientist does it. In larger teams, the data scientist comes to the data engineer / big data engineer and says. "Look, I've come up with a super great model, let's do it like that"
  • Visualization of data analyzes and derivation of recommendations for action for departments
Ne that are rather boring Excel and PP Presis.
  • Providing real-time analytics to decision makers
"Can you tell me quickly how many customers have this and that with x, y, z" -> SQL or Spark queries etc.
  • Execution of maturity level analyzes for BI as well as BI business value analyzes
I don't know either.
  • Methodical guidance in the field of analytics, especially for exploratory and predictive analyzes
Sounds more like training + Jupyther Notebook sheets.

It depends, if you're the only data scientist, of course, you do everything.
Re: which master is right for me? (Data Science / Machine Learning)
From anonymous It depends, if you're the only data scientist, of course, you do everything.
That's right, on Facebook, for example, they only do A / B testing all day.
We don't have any data engineers, I have to get everything myself. Do you have a team? If so, how big is it?
Anonymous  📅 30.06.2017 21:42:28
Re: which master is right for me? (Area of ​​data science / machine learning)
From imperius
From anonymous It depends, if you're the only data scientist, of course, you do everything.
That's right, on Facebook, for example, they only do A / B testing all day.
We don't have any data engineers, I have to get everything myself. Do you have a team? If so, how big is it?
Only 2, Analyst and me. I'm also more of a software developer with an ML focus.
Why this topic ended

Closing a topic happens automatically when that Subject old and there have been no new posts for a long time. The background is that the statements made in the thread may no longer be applicable and it does not make sense to discuss them further.Please inform yourself in newer contributions or in our editorial articles!

Newer topics are sometimes closed by the moderator if it has the feeling that the topic has been discussed or, above all, attracts unpleasant people and / or trolls.

If you still have questions, it is advisable to first see whether there are any out-of-date articles on the respective topic Students online there or whether there may be more current topics in the forum. If that is not the case, you are of course welcome to open a new topic 😇
This forum is operated with the free software Phorum.