Rankings Methodology
Version 1.0. How our “Best” selections are made, what evidence they use, and what they do not use.
Last updated 11 October 2026
Our “Best” and “Top” lists are editorial selections, not official academic rankings. A program is selected when verified facts from its own official pages satisfy every required criterion for that list. Selections are shown alphabetically and are not numbered: we do not have comparable outcome data that would justify ordering one program above another.
Every entry shows the evidence that satisfied each criterion, the strengths and limitations we found in the official information, the kind of applicant the program's published features suit, and the date we reviewed it. Each fact links to its source on the program page.
We do not use acceptance rates, graduation or job-placement rates, salaries, student-satisfaction figures, employer or peer reputation surveys, or institutional rankings — we have not verified comparable figures for these programs, and we do not borrow prestige from a university's other departments. A school-level selection is based only on that school's verified data science degrees.
Completeness is not a proxy for quality. Criteria ask for evidence of specific features — a documented curriculum, a capstone or research component, a stated attendance model, published costs and prerequisites — not for how many fields a profile contains. When evidence is missing, a program is not selected; that does not mean it is weaker, and the list says so.
Sponsorship and advertising never influence selection, order or assessment. Paid placements are displayed separately, labelled “Sponsored”, and the selection code does not read sponsorship data. A school cannot pay to be added, kept, described or ordered.
Applicant-fit labels (for working professionals, for research-oriented applicants, for applicants changing careers…) are used only when official wording supports them, such as a stated part-time option, a thesis track, or admissions that do not require a prior computer science degree.
If numerical rankings are introduced later, we will publish the criteria, weights, data period, missing-data treatment, the universe considered, every score breakdown and a new methodology version before any number appears.
Criteria
| Criterion | What counts as evidence |
|---|---|
| Relevant curriculum coverage | The program's official pages describe its data science curriculum (core courses, required areas or named specializations) in verifiable terms. |
| Capstone, project or research component | A capstone, applied project, thesis or dissertation is documented on an official page. |
| Published structure | Delivery format and program length (duration or required credits) are verified on official pages. |
| Clear attendance requirements (online lists) | For online programs: either official wording supports “fully online” (no required in-person attendance), or each required in-person element (orientation, immersion, residency, proctored exam) is documented. |
| Cost transparency | Tuition is published on an official page with its basis (per credit, per term, per year or whole program); an academic year is shown when the university prints one. |
| Admissions accessibility and clarity | The program publishes its admission requirements — prerequisite expectations (mathematics, statistics or programming) or its GRE policy, plus at least one further requirement. |
| Evidence of funding (PhD lists) | For doctoral selections: an official funding policy is verified — tuition coverage plus a stipend for admitted students (guaranteed or with published conditions). |
| Published financial support | Scholarships, fellowships, assistantships or a funding policy are documented. Not required; recorded when present. |
| Source quality and freshness | Eligibility evidence and the facts above come from the university's own pages and were verified within the last 12 months. |
Faculty and research information is not used as a criterion in version 1.0: comparable, sourced faculty data is not available for most programs, and counting names on a page would reward page length rather than quality. Program-specific outcomes are not used because very few programs publish them on a comparable basis.
The selections and their required criteria
| Selection | Eligible degrees | Required criteria | Minimum |
|---|---|---|---|
| Best Data Science Schools in the USA | Schools with at least three verified, eligible data science degrees (dedicated or joint) across two or more degree levels. Related degrees with only a data science concentration do not count toward a school's inclusion. | At least three verified data science degrees across two or more levels; at least two selected in our program lists | 5 |
| Best Master's in Data Science Programs | Dedicated master's degrees in data science and joint master's degrees whose official title names data science. Related degrees with only a data science concentration are not eligible. | Relevant curriculum coverage; Capstone, project or research component; Published structure; Cost transparency; Admissions accessibility and clarity; Source quality and freshness | 5 |
| Best Online Master's in Data Science Programs | Dedicated and joint data science master's degrees offered fully online or online with required in-person attendance. Hybrid programs with substantial in-person coursework are not eligible for this list. | Clear attendance requirements (online lists); Relevant curriculum coverage; Published structure; Cost transparency; Admissions accessibility and clarity; Source quality and freshness | 3 |
| Best Bachelor's in Data Science Programs | Dedicated bachelor's degrees in data science and joint bachelor's degrees whose official title names data science (for example Statistics and Data Science). Degrees with only a data science concentration are not eligible. | Relevant curriculum coverage; Capstone, project or research component; Published structure; Source quality and freshness | 5 |
| Best PhD Programs in Data Science | Doctoral degrees whose official title names data science (dedicated or joint). | Evidence of funding (PhD lists); Capstone, project or research component; Published structure; Admissions accessibility and clarity; Source quality and freshness | 5 |
A selection with fewer qualifying entries than its minimum is not published. The PhD selection is not published until enough doctoral programs publish a verified funding policy.
Strengths, limitations and applicant fit
Each entry lists strengths and limitations derived from the same verified facts — for example a stated part-time option, a published estimated total tuition, a GRE requirement, or attendance rules that are not published. Applicant-fit labels are used only when official wording supports them:
- For working professionals — a stated part-time option, or an asynchronous online schedule.
- For research-oriented applicants — a documented thesis or dissertation, or a doctoral program.
- For applicants changing careers — admissions wording that accepts applicants from any background or offers bridge or foundation courses.
- For applicants who need a fully online program — the verified Fully online badge.
We do not name a universal “best overall” program: the available evidence does not support one.
Data period, universe and missing data
Data period: Facts verified on official pages in October 2026 and on the review date shown on each list.
Universe: Published, eligible programs in our directory at review time (see /methodology/ for eligibility). Coverage is partial and growing; a program missing from the directory cannot be selected.
Missing data: Unverified facts count as missing. A required criterion with missing evidence means the program is not selected; optional criteria are recorded when present. Missing evidence never counts in a program's favor and is never treated as a negative fact.
Sponsorship
The selection code does not read advertising data. Sponsored placements are labelled, displayed outside the lists, and cannot buy inclusion, order, description or verified status. See the sponsorship disclosure.