How the UR AI Score is calculated

The UR AI Score is a weighted average of six signals, each normalised to a 0-100 scale before weighting. Every university we cover is scored with the same model, and every ranking page you see — country, city, subject, affordability — is that model sorted and filtered, never a hand-picked list. That is the whole mechanism. It is deliberately simple, because a score nobody can reconstruct is a score nobody should trust.

Three weightings exist rather than one. General rankings use the default weighting. Affordability rankings shift most of the weight onto tuition and living cost. Subject rankings shift it onto programme reputation and career outcomes. The weighting in use is printed on each ranking page, so you can see which question the list is actually answering before you read the order.

The score is a comparison aid, not a verdict. It is most useful for narrowing forty candidates to six. It is least useful for choosing between two universities two points apart, where the gap is well inside the error of the inputs — that is the moment to stop reading rankings and start reading fee pages and graduate outcomes, which is the argument made at length in our guide on turning a degree into a first job.

Inputs come from four families: official university pages (programmes, tuition, admissions, student support), national registers and quality-assurance bodies, publicly available outcomes datasets, and labour-market signals about the industries around each campus. Where a specific source underpins a specific page, it is named in the Data sources panel at the foot of that page.

Two honest caveats. First, tuition on this site is an estimated band, not a quotation: published fees move every year, differ by nationality and programme, and rarely include the living costs that dominate an actual budget. Second, a substantial share of the reputation signal ultimately traces back to a small number of large public ranking systems, so the score inherits their known biases towards older, research-heavy, English-speaking institutions.

Refresh cadence differs by input. Programme and tuition pages are re-read quarterly; registers and outcomes datasets annually, because that is how often they change. A page that has not been touched in months is not necessarily stale — it usually means none of its inputs moved.

The score is a filter, not a decision. Use it in the order below and it will save you weeks; use it as a league table and it will make a decision you should be making yourself.