Back to index
Other30 credit hours (10 courses), asynchronous; part-time students commonly take 2 to 3 years·Priced at $10,000 plus fees

UT Austin Online Master of Science in Data Science (MSDS)

4.3

UT Austin's data science sibling to its MSAI, and probably the best balance of statistics and modern ML of any cheap online degree. Five required courses cover the foundations, and the elective list lets you go deep into generative models and NLP. At about $10,000 plus fees it is hard to beat on value.

What We Liked

  • About $10,000 plus fees for a master's from a top statistics and CS school
  • Proper statistics core: probability and inference, regression and predictive modelling
  • ML electives are modern, including deep generative models, deep learning and NLP
  • Shares several courses with UT's MSAI, so the ML teaching is the same quality
  • Asynchronous with weekly deadlines, which works across time zones

What Could Be Better

  • Fees on top of the $10,000 are not clearly totalled on the program page
  • Data Structures and Algorithms is required, which will be a wall for some analysts
  • Smaller community than Georgia Tech, so fewer course reviews to plan with
  • No practicum or capstone, so you need to build your own portfolio project

Detailed review

UT Austin's online MSDS came before the MSAI and uses the same model: ten three-credit courses, asynchronous lectures with weekly deadlines, and a headline price of $10,000 plus fees. It gets about 260 searches a month in the US for ut austin msds, which is small next to Georgia Tech's OMSA, but I think it is one of the most underrated options in this space. Five courses are required. Principles of Data Science is the gateway course in your first semester. Then Data Structures and Algorithms, Machine Learning, Probability and Inference, and Regression and Predictive Modeling.

That is a good core. It covers the statistics that self-taught data scientists usually skip, plus enough CS that you can work on a real engineering team. The five electives are split into two groups. From Statistical Modeling and Decision Analytics you take two, from options including Advanced Predictive Models, Data Exploration and Visualization, Data Science for Health Discovery and Innovation, and Design Principles and Causal Inference. From Machine Learning and AI Systems you take three, from Advances in Deep Generative Models, Advances in Deep Learning, AI in Healthcare, Deep Learning, NLP, Optimization and Reinforcement Learning.

Many of the ML electives are shared with the MSAI, so you are getting the same courses for the same price. Application dates match the MSAI. Fall opens on 15 December with a priority deadline of 15 March and a final deadline of 15 April. Spring opens 1 June, with 1 August priority and 1 September final. There is also a shorter graduate certificate if you want to test the waters. The weaknesses are minor.

The program page says $10,000 plus fees without adding the fees up, which I find annoying. There is no capstone, so you have to make your own portfolio work. And requiring Data Structures and Algorithms will put off some analysts, although I think it is the right call. Overall, for someone who wants a rigorous, modern data science degree without debt, this is one of the two or three best choices available.

[ final ]

The verdict.

My pick for someone who wants real statistics plus modern ML and does not want to pay more than about $10,000. Choose this over Georgia Tech's OMSA if you want more ML and less business. Choose UT's MSAI instead if you are sure you only care about AI and never want to see a regression table again.