MADS is a fully online master's from Michigan's School of Information. It takes 38 credits, mostly delivered as one-credit, four-week courses, and most people finish in 12 to 36 months. There is also a version listed on Coursera, but the Coursera pages I checked were showing older tuition and deadlines, so I would go straight to the UMSI site. For 2025-26 the School of Information lists $1,161 per credit hour for Michigan residents and $1,548 for everyone else. That works out at roughly $44,000 and $59,000 for the whole degree before the small per-semester fees.
Scholarships of $1,000 to $10,000 exist, with priority given to people who apply by the early deadline. Admissions are deliberately easy. There are no essays, no letters and no GRE, and from fall 2026 there is no application fee. Admitted students take a programming placement test. I like that.
It assumes people can prove themselves in the program rather than on paper. The content is applied and sensible: data manipulation, visualisation, ML, NLP, ethics and a portfolio project at the end. The four-week format keeps you moving, but I found the shorter format squeezes the harder topics. Machine learning in a month is a sprint, and you will want to do extra reading. Search interest is small, around 20 to 30 US searches a month for university of michigan mads. The people looking are mostly already decided on Michigan. My view: it is a good degree with a price problem.
The UMSI angle on ethics and human-centred data is a real differentiator, but for most self-funded learners the cheaper Texas and Georgia programs are the better deal.