[pj_mukh]

Maybe they could've focussed on making the credentials valuable (maybe they did and it didn't work). AFAIK, if you told an employer you have a "nanodegree" in something or other, its glossed over.

Usually, you have to back it up with real-world projects. In which case why would I pay for the degree, I should just go out and build something with the stuff I learn for free.

[strictnein]

Some NSA job listings have mentioned Coursera's (and others) Data Science specializations.

> "Completion of a data science certificate program (online or other) may replace 1 year of relevant experience. Some examples of data science certificate programs include those offered by Cloudera, Coursera, Indiana University-Bloomington, and University of California-Irvine."

Data Scientist - Entry Level: https://www.nsa.gov/psp/applyonline/EMPLOYEE/HRMS/c/HRS_HRAM...

Data Scientist: https://www.nsa.gov/psp/applyonline/EMPLOYEE/HRMS/c/HRS_HRAM...

Might just be an indication of how desperate they are for those positions, but it could also indicate they're doing a pretty good job educating people.

[KKKKkkkk1]

The US government has a lot of trouble hiring qualified scientists in the high-end disciplines because (a) It is not willing to pay market salaries; and (b) It artificially restricts the candidate pool to US citizens who are able (and willing) to pass security clearance. So I wouldn't be surprised if they'd be hiring data scientists straight out of community college.

[Animats]

NSA has a big in-house training operation. Unlike startups, they're not looking for narrow specific skills in new hires. A year of paid formal classroom training for a new hire is not unusual at NSA.

[intrasight]

I looked at their Data Science degree (am still looking). It appears to be a quality offering.

[Programmatic]

I began the JHU Data Scientist Specialization on Coursera and worked through the Data Scientists Toolbox and R Programming classes, but found them a bit lackluster in their presentation.

The course content (videos, slides) presents some basics with a large emphasis on having a "hacker ethos" to do more work and digging on your own to be successful. The quizzes and projects in the R course demanded the use of techniques that were outside of the content provided in the lecture. I am not against having a hacker ethos and personally am happy to research and learn on my own, but I fail to see the logic in charging money to tell people to Google. I expect a curriculum to be a self-contained unit of learning, or else I wouldn't bother with curriculum.

Contrast this to Coursera's excellent (and free) Rice University Interactive Python Programming classes that are really superb. The Rice University team put together an online Python interpreter complete with graphics capability so that students could test each other's code. The R class left peer review of students' code as "do not run the code, eyeball it and see if it looks right". I understand why they did it (execution of potentially harmful code), but Rice's solution was elegant.

[mindcrime]

I am not against having a hacker ethos and personally am happy to research and learn on my own, but I fail to see the logic in charging money to tell people to Google.

I'm about halfway through the JHU Data Science program and I agree that there's a fair amount of "extra" work you have to do. But still, coming into it as somebody who had never used R at all before, I've learned a ton and have definitely found it to be worth the money I've spent (I plan to actually do the entire track and get the certificate and everything).

Now that said, I probably woulnd't pay much more than what they're charging now (I think most of these classes have been $49.00 / class so far), but at that rate, I definitely feel that it's been worthwhile. And even if it isn't a formal degree, it still gives me the ability to legitimately mention JHU on my resume. So if somebody is just quick skimming my resume to make a decision on which pile to sort it into, that probably gives mine a small nudge in favor of the "look deeper" pile.

[ghaff]

Oh, I'm pretty sure they were all aware that credential value was ultimately going to have to be a pretty important component of getting people to pay more than a token amount. But the value of credentials is much more something that others confer on the credential than something the credentialing organization can confer on itself.

Sure, the organization can attempt to explain and promote how its credential provides evidence of X, Y, Z skills or knowledge. Which works reasonably for both well-established institutions and narrow vocational training (including in IT). But it's a lot harder to make a case that random MOOC course provides evidence of much other than a generalized interest in continuing education (for which you don't really need a credential to prove).

[coredog64]

I started on a Udacity cert program, but had to drop it due to time constraints. The capstone of the program is/was a real-world project.

While I never signed up for the nanodegree track, it appears those are similar.