Looking over the list of ACS 35k+ variables available, it's hard to narrow down on exactly what I'm looking for. Is there a tool or set of best practices available? For example, if I wanted to find household income tier count by householder age and presence of children, how would one go about that?
I've not found a better way than other than to find a survey section and just scroll.
TIA!
I use the first two numbers after the B or C to narrow things down. For example, all income tables start with 19, and earnings tables start with 20. So it would be tables that look like this: B19001, B20001, etc.
Expand Element 2: Subject for the list here: https://www.census.gov/programs-surveys/acs/data/data-tables/table-ids-explained.html
Thanks Diana.
This is kind of what I needed up doing, but without the excellent reference you linked. I'm surprised there's not a simple tool somewhere where you can select topics and filter everything down to specific variables.
I may have to work on that some time.
RE: "a simple tool somewhere where you can select topics and filter everything down to specific variables", I believe that's what data.census.gov is intended to be, at least in its "Advanced Search" mode. It's also what IPUMS NHGIS (which I work on) is designed to do. It's hard to design a tool that's both "simple" and "robust," which is why I expect you'll have reason to quibble with both these options, but they're a good place to start if you haven't tried them yet.
Is any organization experimenting with using AI/LLMs to develop a semantic type search for Census variables?
I hear ya. I don’t need help munging the data. That’s the easy part for me… I prefer to load via the api in a long format. I just need help finding the best variables to use. Seems that will just come in time.