Every dataset is coded to the Standard Occupation Classification (the 2010 version).
This basically means that if you want to get information about any job, you will have to look up that job's SOC code first.
The API provides a facility for this under the /soc endpoint.
The Annual Survey of Hours and Earnings (ASHE) is a very comprehensive and detailed dataset. It's not available to the public for privacy reasons.
When you ask the API for ASHE data, you actually get 'non-disclosive' data that's been mangled – just a little – to protect people's privacy. 1)
This means that if you ask how much a 46-year old female dentist in Wales gets paid, the result will not be what she actually gets paid, but a number that is very much like what she would get paid. If there's only one person like that, the number will be an average between persons very much like her, so you can never pinpoint exactly how much any one person gets.
While our data team works on figuring out the best way to make this data non-disclosive, the ASHE API will give you data that's sometimes more and sometimes less accurate.
In summary: ASHE results are very good estimates, but not the actual facts. They will also be fairly unreliable until the API is publicly released.
While the full O*NET set provides a wealth of information, the current API only delivers info about STEM2) jobs and skills..
Employer Skills Survey (ESS)
Actually you get the result of an estimation function that's been run against the original data.