Choosing a Data Science PhD

A doctorate with "data science" in its name can mean several different things. It can be a PhD in Data Science run by a dedicated school or institute, or a joint PhD that pairs data science with statistics or computing. It can also be an established computer science or information science PhD with a data science concentration, or a professional doctorate that is not a PhD at all. In our directory (October 2026) we list 22 doctoral programs (how we count), and they differ in structure, length and funding as much as in name. This guide covers what to compare before you apply.

Three kinds of research doctorate

PhD in Data Science. The degree is named for the field and usually run by a data science school, institute or department. Examples include the University of Virginia's School of Data Science, NYU's Center for Data Science, UC San Diego's Halıcıoğlu Data Science Institute, Stony Brook (run jointly by Applied Mathematics and Statistics and Computer Science), the University of North Texas and Worcester Polytechnic Institute.

Joint-field PhDs. Here data science is paired with a parent discipline in the degree title. Examples are the PhD in Statistics and Data Science at Northwestern, at Penn's Wharton School and at the University of Arizona (an interdisciplinary program), and Boston University's PhD in Computing & Data Sciences. These programs often carry the theoretical expectations of the parent field. Arizona, for example, requires a Ph.D. minor.

Related PhDs with a data science concentration. The degree you earn is in another field. Florida Atlantic University's PhD in Computer Science with a Data Science and Analytics concentration requires at least four graduate courses from a data science list and a dissertation with "a strong emphasis on one or more areas of data science and analytics." It's still a computer science PhD, and you should judge it on that department's strengths.

None of these is automatically better. The right choice depends on the kind of research you want to do and the faculty who do it.

Professional doctorates are not PhDs

Some data science doctorates are applied, professional degrees. Lewis University's Doctor of Engineering (D.Eng.) in Data Science gives students "experience through a long-term applied project". Students must pass a qualifying examination, get a project proposal accepted and "defend their project in front of a committee." The catalog requires at least 84 credit hours without a prior master's (up to 93 with foundation courses), or 54 with a related master's, and says the program can be completed on-ground or online.

A D.Eng. can suit a working professional who wants a doctoral-level applied credential. It is a different degree from a research PhD, though, and it isn't the usual route into academic research careers. Read the program's own description of its degree type before you treat it as a PhD alternative.

How programs are structured

Most research PhDs follow a similar arc: coursework, a qualifying or comprehensive exam, a proposal or prospectus, then dissertation research and a defense. The details differ.

• Virginia: typically five years of full-time, in-person study for students without a related master's. Coursework fills years 1 and 2, and year 3 is a transition to independent research. Students usually spend the first summer in one or more faculty labs, and the comprehensive exam is a multi-part, in-person exam.
• NYU: a research rotation taken six times (18 credits), a Comprehensive Exam, a Depth Qualifying Exam and a teaching requirement.
• UC San Diego: a research rotation program and a preliminary assessment examination, followed by a dissertation defense before a committee of four or more members.
• Northwestern: a comprehensive exam typically taken in the fall quarter of the second year, then a prospectus and admission to candidacy by the end of year 3.
• Stony Brook: a written qualifying exam made up of a 3-hour CS foundation exam and a 3-hour data mining exam.
• Boston University: research rotations, an oral area exam, a dissertation prospectus and a final thesis defense.
• Kennesaw State: lets you "choose from either the research model or the practitioner model."

Rotations matter if you're unsure which lab you want to join. Programs without them generally expect you to identify an adviser sooner.

Credit requirements aren't comparable unit for unit

Published totals vary widely, partly because schools count in different units (semester credits, quarter units, Penn's course units):

Credit totals also tell you how a program treats a prior master's. Boston University and North Texas publish lower totals for master's holders. WPI asks students who enter with only a bachelor's degree to first complete its MS in data science using the thesis option. Virginia, by contrast, says a master's degree is not required.

Research fit and advisers

The program name matters less than the people you'd work with. Before applying:

• Read recent papers by the faculty you'd want as advisers, and check whether they are taking students.
• Find out whether admission is to the program or to a specific lab, and how advisers are matched.
• Decide whether you want methods research (statistical theory, machine learning) or applied research in a domain. Joint statistics programs and applied programs lean in different directions.
• Ask current students about time to degree, teaching load and how often students change advisers.

Funding: read the exact wording

Funding terms vary as much as structure. Virginia states that admitted applicants receive support for the duration of enrollment, subject to continued enrollment and satisfactory progress, and that it only admits funded students. Boston University guarantees PhD students in good standing a five-year funding offer. At Wharton, all admitted doctoral students receive a fellowship covering tuition and health insurance plus a stipend. Kennesaw State says "qualified students" receive a tuition waiver and a research stipend, which is not the same as funding for everyone admitted.

Understanding PhD Funding in Data Science explains how to read these statements. Our funded PhD list shows the programs whose published policy we could verify. You can browse every doctoral program on the PhD page.