ACS Data Users Group

Hi Group,

We're trying to figure out if we can query 2015 ACS 5-year estimates for all block groups within a state for certain IDs/variables. An example screenshot of what this would look like in American Fact Finder (AFF) is attached. Right now it looks like 2015 ACS 5-year estimates are only available through the API (and not FTP), and we're not sure if it is possible to request all block groups for an entire state at once as we are getting an error. Does anyone know if this request is even possible through the Census API for ACS?

Thanks,

Tom

Hey Tom,

I've not had any luck downloading BG data for entire states directly from American FactFinder without the use of an API. What I've used instead with great success is NHGIS, (https://www.nhgis.org/), which hosts very well-organized tabular and shapefile data. You can download BG data for an entire state or US if you wish.

Thanks Ian for the info.

We are trying to figure out if there is an efficient way to query 2015 ACS 5-year estimates for all block groups within a state using an API or FTP approach. Right now, it looks like only API is available and we get an error when we submit the request for all block groups within a state. We're trying to figure out if this is even possible through the API.

I have looked at NHGIS before. My only concern with NHGIS is their stricter data use agreement and the process is still manual (at least I didn't see an API or FTP through NHGIS).

Oh I see. Yeah it is manual, though easy to select by table ID. But if this isn't a one-off research project, I'm not sure. I've tried unsuccessfully before to use the ACS package in R, which takes advantage of the census API to download data. I read that there is a way to nest download calls so that you can get data for large geographies like states, but wasn't able to do it on my own.

cran.r-project.org/.../acs.pdf

Very common complaint about the ACS data that I would also like a solution to ...

Yeah, I wish it was a little more clear. I should have known, there is an R package for this. This might help us out. :)

NHGIS might be a good short term solution; though, I'm still trying to figure out the value add of NHGIS and/or if we back ourselves into a corner in some way by using there processed ACS data.