Taking this with a huge grain of salt (like any study), this is still an interesting look just for it's rarity.
Short version of the story: an Obama era program called Gainful Employment tracked university programs that qualified for federal student aid, federal loans in particular. The big data points were the institution, degree program, debt taken on to achieve that program, and first year earnings after graduation.
There's only two years worth of data, for the 2014-2015 and 2015-2016 academic years. The Trump administration got rid of Gainful Employment regulations. Why depends on who you ask, but it looks like the penalties targeted for-profit institutions while giving public and non-profit institutions a pass.
But even if the regulations are gone, the data remains. We now have a snapshot in the recent past of about 40K different college programs, the mean and median debts taken on to graduate in those programs, and the median annual earnings of someone who graduated in it.
From there, the program extracted two more points.
One, Debt to Earnings ratio. Effectively, debt as a percentage of annual income. (A lower number here is better)
Two, Gainful Employment Status. The old regulations that this data was set for put programs in three categories, based on whether or not the debt incurred was a reasonable percentage of earnings. The categories were pass, fail, and zone (a kind of probationary in-between segment). The math on how this was calculated, along with ways to adjust on your own if you wish, is included in the data table.
This data doesn't answer a question like, "is a degree worth it?"
This data calculates odds on institutions, degree programs, debt levels, and known earnings to let individuals decide.
And to my knowledge, this is the only dataset of it's kind that actually held university programs accountable to the future earnings of its pupils.
It's not comprehensive. It's not easy, or simple, or definitive.
But it's better than anything else I've ever seen.
(Check it out)