Eric Potash

Thanks. Is the dataset I linked to what you refer to as the first or second iteration?

I have written a [script](https://github.com/potash/census-postgres/blob/master/meta-scripts/build_import_sql.sh) that creates an "import.sql" script for ACS 5-year data (it is currently hardcoded for illinois but should work for the country). For years after 2010...

Huh, can you say more? Maybe I wasn't clear. The weighted dataset only contains the unique observations `x,y` and an additional `n` indicating the number of such observations in the...

Another hint to me that it's not working is that removing the weights from ```r stan_glm(y ~ x, data=df_weighted, weights=n, iter=500) ``` to ```r stan_glm(y ~ x, data=df_weighted, iter=500) ```...

That's my expectation as well but it's not the result I'm getting

Above `df` is uncollapsed, `df_weighted` is collapsed

@bgoodri when you get a chance can you take a closer look? I'm pretty sure that the behavior you are saying should happen is not happening. Thanks

Got it thanks for the update. Hope the flux gets stabilized

Btw I added a fake group to the data and can confirm that weights in `stan_glmer()` have the desired result, i.e. uncollapsed and collapsed weighted fits agree, collapsed unweighted fit...

A couple questions before you shut this down: 0) When going through ag data commons today to try to download RaCA I get pointed [here](https://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/research/?cid=nrcs142p2_054164) which is broken. What is...