Glassdoor through HasData
Two tools. hasdata_glassdoor_listing_getJobListings takes keyword and location and returns thirty postings a page. hasdata_glassdoor_job_getJobDetails takes the posting url and returns that one posting in full.
The listing already carries the salary, the employer rating, the skills and the age of the posting, so most comp and hiring questions are answerable without opening anything.
The salary field is two different things
salary.source reads either EMPLOYER_PROVIDED or ESTIMATED. The first is a figure the employer published, the second is Glassdoor's own model for that role and market. Averaging across both produces a number that describes neither, so split on source before aggregating and name which set a published figure came from.
min, median and max can all be equal, which is a single stated salary rather than a broken range. salary can also be absent, and period can read HOURLY, so converting to an annual figure is an assumption you are making rather than something the response did.
What only the detail call has
The listing description is a snippet and it ends mid-sentence. Full text, stated qualifications and the expired flag live on the detail tool, and a listing page still shows postings the employer has closed. Reading every posting in full costs one call each, so open the few that matter rather than the page.
Paging and domains
pagination.otherPages maps a page number to an opaque token, and that token is what nextPageToken takes. Keep the other parameters unchanged while paging.
domain selects one of the regional Glassdoor sites, and the US site is the default rather than a value you can pass.
Salary figures and employer ratings are Glassdoor's own numbers. Attribute them instead of presenting them as the market.
