Which master's is better for a position in data science - CMU computer science/machine learning vs. Harvard computational science engineering?

Machine Learning is a better avenue because many of the concepts that industry requires of Data Science is predictions. CMU is also a very well known CS school (with a stronger program than Harvard), so your chances of being hired in Silicon Valley and running a team would be higher. I should note that NYU's Data Science program is working in collaboration with Berkeley and a few other universities on streamlining the academic concept of Data Science, so while the name may not have the wow factor of CMU and Harvard, it could be part of a ground breaking evolution in the field.

I don’t know the program from the inside, but Harvard’s program does not seem to be what you want. Computational science and data science are very different fields. Computational science is exactly what it sounds like: computational methods use to perform scientific analysis, like fluid dynamics and genomics, and data tend to be structured, uniform, and homogeneous. Data science is really a mix of CS, stats, ML, and analytics, and is used by businesses (typically) and other organizations to reason about and over large sets of data that are usually heterogeneous and a mix of structured and unstructured.Chicago’s stats program is not going to be as cutting edge or employment-friendly as the NYU or CMU programs and is likely to be more theoretical, but they’re probably not vastly different.However, let me make a slightly different argument for CMU. Pittsburgh is livable and affordable. For the year I was in Pittsburgh, I shared the first floor of a ;arge Victorian (we only shared a kitchen, as we each had our own enormous bedroom and our own bathroom) just a mile north of CMU’s campus for around $400/month (this was in 2009). I had a car and ready access to parking, and got to enjoy life (including access to excellent and affordable cultural activities). All of my friends who did their PhDs at Pitt and CMU lived comfortably. In comparison, last summer I was paying $1,000/month to sublet half of a nice but cramped Cambridge two-bedroom apartment from a Harvard Law student, and that was hundreds cheaper than my prior shared apartment in Jamaica Plain. Having a car was impossible, both because of the expense and lack of places to park, so other things were expensive, like food (and a PITA to shop for). I have never lived in NYC, but I’ve lived near. Manhattan is just unaffordable for grad students who are not independently wealthy. Neighborhoods are gentrifying rapidly, so what little affordable housing there is in the outer boroughs is fought over by long-term and short-term residents.It would make sense to take on the additional debt and deal with frustrating living circumstances for a couple of years if the program were superior, but that’s not the case here. It may lack the broader reputations enjoyed by the other schools on your list, but computer science (and anything else computational, by extension) is the living, breathing heart of CMU, and there is no better CS program (whether others are equals is a topic for another question that I would avoid like the plague). In industry, CMU students, especially graduates of CMU’s specialized computer science master’s programs (e.g., software engineering, HCI) are able to write their own tickets. Stanford’s MS CS ekes out a slight advantage only because of the right-there access to Silicon Valley and venture capital.