Back to index
Other30 one-credit courses in 8-week sessions; about 24 months at three courses a session, up to 8 years allowed·$15,750 total ($525 per credit hour), paid per course

CU Boulder Online Master of Science in Data Science (MS-DS)

4.0

The most open door of any data science master's I have looked at. There is no application. You take three for-credit courses, get a 3.0 or better, and you are in. The trade-off is that the degree is built from 30 small one-credit Coursera courses, and some of them feel more like a MOOC with a grade attached than a graduate class.

What We Liked

  • Performance-based admission, so no transcripts, essays or recommendation letters
  • No application fee and you only pay for the courses you take that term
  • $15,750 total is fair for a degree from a solid research university
  • Eight-week sessions make it easy to pause for a busy month at work
  • Courses overlap with Coursera specialisations, so you can audit before paying

What Could Be Better

  • 30 one-credit courses means a lot of thin, quiz-heavy units rather than deep projects
  • Much less instructor contact than a cohort program
  • Search interest for the degree is down about 44 percent year on year, which suggests the Coursera degree market is cooling
  • Proctored online finals in some courses, which some people hate

Detailed review

CU Boulder's MS-DS runs on Coursera and is made up of 30 one-credit courses, each taking about eight weeks and around 4 to 6 hours a week. The total is $15,750, charged at $525 per credit hour as you go. Take three courses each session and you can finish in about 24 months. You have up to eight years. The admission model is the reason to look at it.

There is no application and no fee. You enrol in the for-credit version of courses, and once you have passed three with a 3.0 or higher you are admitted. Boulder recommends some Python, R, calculus and linear algebra first, but does not require proof of any of it. For career changers with an old or weak transcript, this matters more than anything else on the page. The downside is the shape of the courses.

One credit each is small. Many feel like a Coursera specialisation module with stricter grading, and the depth in machine learning is lighter than a three-credit graduate class at Georgia Tech or UT Austin. You also have less contact with faculty, and the peer community is spread thin across rolling cohorts. On search demand, cu boulder ms data science gets roughly 40 US searches a month, down about 44 percent on last year. I would not read too much into that, but it fits a wider pattern of online degree searches dropping as people turn to short AI courses instead. My honest take: it is a fair price for a legitimate degree, and the open admission is genuinely good for people who were filtered out by traditional applications.

Just plan to build your own serious projects alongside it, because the coursework alone will not make you stand out.

[ final ]

The verdict.

Good for someone without a strong academic record who wants a real master's and is disciplined enough to self-pace. If you can get into Georgia Tech's OMSA or UT Austin's MSDS, I think those give more depth per dollar. If you would struggle to get admitted anywhere, Boulder is the honest way in.