[Imports] United States Poultry Import

Spencer Alves impiaaa at gmail.com
Wed Apr 20 00:51:18 UTC 2022


The talk page on the import notes that the features to be imported are 
actually heterogeneous—even though you're tagging all of them with 
man_made=works, the features could represent anything that processes 
poultry, including supermarkets, delis, butcher shops, fast-food 
restaurants, and possibly more. I don't think it's appropriate to import 
all of these features under the same tag. I think that part of the 
import process should include classifying these correctly.

The import page says that "the data size is way too small to warrant a 
team," however nearly 4,500 features spread across multiple continents 
sounds like it would take a while for one person to review entirely, and 
the features should all be manually reviewed if they all need to be 
properly classified. Manual review would also allow for the 
man_made=works points to be upgraded to areas where appropriate.

The page seems to indicate that the import will not be split into 
multiple changesets. Local mappers appreciate when changeset bounds only 
cover a local area. Even if you decide to do all of the importing 
yourself, consider batching them by city, ZIP, county, or state.

I generally disagree with adding source tags on features themselves, 
rather than just the changeset, but if you do insist on importing 
unreviewed data then I suppose it will help with cleanup. However, the 
import page does not specify any procedure for cleanup after the import. 
If the data is not manually reviewed and bad data does make it onto the 
map, the import also needs to include a procedure for cleanup.

It's great that the data includes addresses, however it looks like the 
parsing code goes wrong in some places. For example, addr:street 
sometimes includes part of the house number. I'd suggest making a step 
in your process check if addr:street matches the name of a street 
nearby. They don't always have to match (one example I found has 
addr:street=State Highway 31, which is correct, despite that street not 
having a name), but this step can help find errors.

The "TEXT35temp" field has been left in the sample data. It needs to be 
removed.

Please include details about the data source on the wiki page. The 
ArcGIS link may go stale one day, but there's no other way to tell that 
the data is from the DHS. The wiki page should also mention that the 
data is Public Domain because it was published by the DHS.

On 4/19/22 3:10 AM, Hiausirg wrote:
> I'm planning a relatively small import of chicken slaughterhouses and processing sites in the US - All details are at https://wiki.openstreetmap.org/wiki/United_States_Poultry_Import - see also the talk page there.
>
> Greetings
>
> _______________________________________________
> Imports mailing list
> Imports at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/imports



More information about the Imports mailing list