4 ms·

Ask HN: Thoughts on an MS in Data Science?

3 points by misframer 13y ago

My university just announced an 11-month professional masters program for data science: http://people.virginia.edu/~jjh2b/msds/ Is it worth it?

ilovefood 13y ago
I took one, and I learned more on internet than these sites. If you're commited in learning, there is already everything online for free and you will quickly "get the hands in the dirt", which means actually doing what is interesting. So, imho, this subject moves so fast, that by 11 months you'll have to learn new things.
- urlwolf 13y ago
  Agreed. The online material is outstanding. What worries me is how many people went through it and still feel their data science skills are lacking (90% of people I talked to). 'Yes, I took the Andrew Ng course; it's great, I feel like I learned a lot. But... now what?' Data science is such a complex skillset that to reach mastery nothing beats an apprenticeship (sitting elbow-to-elbow with people who are world-class at it). This is why we are launching an an intensive 12-week pair-programming-based course called Data Science Retreat: http://datascienceretreat.com http://datascienceretreat.com It assumes people have gone through at least one of the popular MOOCs, and know one programming language well. Then it builds from there. On top of that, it tries to bump up business and communication skills, which are often lacking in many PhD-level people. We want to record presentations, given in front of a non-technical audience, and use the recording to give laser-targeted feedback. If you cannot convince the CEO in a 10-min. meeting, all your predictive modeling efforts will be wasted. Disclaimer: I'm one of the organizers.
eshvk 13y ago
So if you actually have those pre-reqs upto scratch: Linear Algebra (up to the level of Strang/Sheldon Axler), Statistics (any undergrad book), Algorithms/Data Structures. You can ramp up on your own time and read key research papers. I would do it only if there are amazing people in UVa who you want to work with.

3 points by misframer 13y ago

My university just announced an 11-month professional masters program for data science: http://people.virginia.edu/~jjh2b/msds/ Is it worth it?

ilovefood 13y ago

I took one, and I learned more on internet than these sites. If you're commited in learning, there is already everything online for free and you will quickly "get the hands in the dirt", which means actually doing what is interesting. So, imho, this subject moves so fast, that by 11 months you'll have to learn new things.

urlwolf 13y ago
Agreed. The online material is outstanding. What worries me is how many people went through it and still feel their data science skills are lacking (90% of people I talked to). 'Yes, I took the Andrew Ng course; it's great, I feel like I learned a lot. But... now what?' Data science is such a complex skillset that to reach mastery nothing beats an apprenticeship (sitting elbow-to-elbow with people who are world-class at it). This is why we are launching an an intensive 12-week pair-programming-based course called Data Science Retreat: http://datascienceretreat.com http://datascienceretreat.com It assumes people have gone through at least one of the popular MOOCs, and know one programming language well. Then it builds from there. On top of that, it tries to bump up business and communication skills, which are often lacking in many PhD-level people. We want to record presentations, given in front of a non-technical audience, and use the recording to give laser-targeted feedback. If you cannot convince the CEO in a 10-min. meeting, all your predictive modeling efforts will be wasted. Disclaimer: I'm one of the organizers.

eshvk 13y ago

So if you actually have those pre-reqs upto scratch: Linear Algebra (up to the level of Strang/Sheldon Axler), Statistics (any undergrad book), Algorithms/Data Structures. You can ramp up on your own time and read key research papers. I would do it only if there are amazing people in UVa who you want to work with.