Welp, failed college and now it's time for a few years of wage slavery. Disappointed in myself...

Welp, failed college and now it's time for a few years of wage slavery. Disappointed in myself, but am not out of the game mentally. It was completely my fault; no effort given, ate cheeseburgers all day with loan money. I do have some plans to get back on track, and if I keep my focus things should be alright.

I did pick up an interest in stats while at school and reached what some schools call intermediate level stats. Lately I've been hearing of people becoming 'data scientists' or 'data jockies', and apparently there are some certs you can get to show your skill at modeling and organizing data in ways that'd be useful to industry. I only know of the intro data scientist cert through Coursera that's, I think, sponsored by Johns Hopkins.

Any current data scientists working in industry currently? Does the cert I mentioned ring any bells? Any information or gfx you guys can provide to help me on my way?

DATA SCIENTIST INFORMATION THREAD

Other urls found in this thread:

r-bloggers.com/in-depth-introduction-to-machine-learning-in-15-hours-of-expert-videos/)
twitter.com/SFWRedditVideos

you're living in the US right?

I don't have anything of value to contribute to this page.

Not unless bumps are valuable? Maybe a bait image? No need to thank me.

The data scientist cert through Coursera you mention, it's not bad, these are good courses but they are intended for people with almost no or very basic knowledge of programming, statistics, etc. There is other online material at places such as datacamp, but they also don't get into advanced stuff, they will teach you a couple of nice things, but they won't teach you how to get the job done, aka how to approach problems and so on.

Want to be a data scientist? Pick up a CS and Statistics Major. Actual data scientists are very well paid, but if you can't even put some effort for college I doubt you can acquire the skills needed for a 150k/year job.

Good references to learn 'data science' are 'Introduction to Statistical Computing: Applications with R' (which the authors provide had done an MMOC course using that book as a reference: r-bloggers.com/in-depth-introduction-to-machine-learning-in-15-hours-of-expert-videos/) and 'Elements of Statistical Learning', a legendary reference in the field.

Yessir

So, more than anything, the certs are just a primer for a big boy diploma?

Basically yes, they are a good introduction to the field. It would be preposterous to think that they are going to make someone a-ready-to-employ-data-scientist.

Are there no such things as higher level certs that cover the advanced topics exclusive to university?

Not that I know of, sorry.

For someone that can't have access, basically due to financial issues, to a higher learning institution, a good start would be the MMOC course I referenced and to progress from there on with textbooks (look up the syllabus of the math and econometrics subjects at the MIT OCW, etc).

If for you learning is not enough, you want to gent some accreditation, a way to do that would be to produce a very good piece of 'data science' project or piece of research with the acquired skills and try to publish it in an academic journal, if possible. This is the closest you can get to obtaining 'certified' skills without going through college (but think that undergraduates don't usually produce publishable research so it's a very difficult path). Presenting a very good piece of work that shows high knowledge of a series of skills would be a good thing to show to a potential employer.

Have you checked at distance universities?

Don't do it. Go do programming instead, easier to get into, and you don't need a degree. Learn a C language, and go make a portfolio of projects.

Take (insert MOOC class here) and do some kaggle competitions. No, not kegel, although that couldn't hurt either.

The fact that they want to represent the data scientist by a smug groomed better-off-homosexual rustles me

That seems pretty accurate to me. They're like engineers but bi.

star a portfolio, read some machine learning books. I recommend machine learning in action. Then shill yourself to companies.

protip: if you fuck up in college explain it to employers by saying a non existent family member died.

There are TONS of certificate programs in statistics. I don't know how possible this is, but I wonder if you could just enroll in a certificate program at a good university in statistics or a related field, like data analysis. Then make sure you do well in those courses, and then off the strength of your grades in those courses reenroll in a BS program, or, at that point, maybe even an accelerated MS

I'd recommend you a Kaggle after a while. It's a real opportunity for you to get a job without degree in math/CS. I'd also recommend CS50 course in Stanford which is free online and very good headstart for understanding computer science. BTW it's not hard course.
I believe that for a sell paid job you do not need to be very smart in science so good luck.

That's an interesting point. I will look into that.

There aren't even good data science programs yet, you have to piece together yourself at Uni.

how does one become a data scientist? i'm almost done with cs/math degree. would i have to get a business bs too? higher than a bs in any of those 3? and i fucked if i haven't already done internships related to it?

Don't you need a statistic degree too?

> Failed college
No offence, but probably not for you. If you're interested in stats and machine learning, learn that. If you're interested in programming, get a job programming. If you're interested in databases, get a job doing something with databases.
If you're looking to do data science for "muh sexiest job of the 21st century" - Harvard Meme Review, and for that mad bank; then expect either the bubble to burst fairly soon when businesses realise everyone wants to do it for the meme and the salaries, and don't have the ability, or expect the field to become better defined and weed out all the people who aren't data scientists, and cause you a lot of problems because you might not be capable of being the ideal data scientist.
If you wanna do data science properly, you'll need to be good at stats, machine learning, programming, databases / business intelligence all at the same time; which short of a few proper college courses in data science (which I assume is not an option to you now) or years of experience (by which time the meme may have died down), you won't be able to do.
There's no shame in working on getting a safe job in programming, and studying machine learning on the side. Maybe you can even get an analyst role doing business inelligence and making simple visualisations, and pass it off for data science experience. MOOCs won't help much.

Yes, you need three bachelor's degrees and at least one further graduate degree in addition to 3+ internship experiences in order to unlock the data scientist entry exam.

>ate cheeseburgers with loans
This is the real reason college debt is a problem in the US.

I bet you managed to fail out of a CS degree. Disgrace.