The problem in doing this is that saying "school X is great for getting a Ph.D." will never likely generally be true, because the Ph.D. process is so advisor dependent. However, doing this for each advisor is really hard too, because the sample size is small and it's hard for the reviewer to remain anonymous.
All of that said, some set of general statistics (size of program, distribution of research group sizes, percentage of those who wanted a Ph.D. and then got one, etc.) plus a general satisfaction survey among graduate students might be a good complement to the usual ratings one sees in US News.
Underrated comment. The “school” rating is meaningless. Even the “department” is meaningless. It all comes down to your advisor.
Our department had amazing advisors and horrible advisors. Two people could be in the same PhD program and have polar opposite experiences.
In fact, if you want reviews it’s not hard - just ask current students about other advisors. Rumors move fast and I’ve found you’ll get an honest opinion.
All of that said, some set of general statistics (size of program, distribution of research group sizes, percentage of those who wanted a Ph.D. and then got one, etc.) plus a general satisfaction survey among graduate students might be a good complement to the usual ratings one sees in US News.