<div dir="ltr"><div dir="ltr"><div>:-) Thank you for your retrospective.</div><div><br></div><div>Please help, <b>How we can improve the import guideline, </b></div><div>AND</div><div>- find & fix the "root cause" ( <a href="https://en.wikipedia.org/wiki/Root_cause_analysis">https://en.wikipedia.org/wiki/Root_cause_analysis</a> )</div><div>- creating a "checklist" for the importers. </div><div><br></div><div>My question is:</div><div>- Which QGIS tool(s) did you use for the conversion? ( version, name, ... )</div><div> probably the address module is not perfect, and we need to report to the maintainers.</div><div> </div><div> example: <font face="monospace"> ADDRESS="45-10 19TH AVENUE" --> <font color="#0000ff">"addr:street" "-10 19th Avenue" </font></font></div><div> </div><div>(later)</div><div>If you can summarize your "lesson learned" ( <a href="https://en.wikipedia.org/wiki/Lessons_learned">https://en.wikipedia.org/wiki/Lessons_learned</a> ) please create a wiki page and link to the <a href="https://wiki.openstreetmap.org/wiki/Import/Past_Problems">https://wiki.openstreetmap.org/wiki/Import/Past_Problems</a> </div><div><br></div><div>(lesson learned for me) </div><div>Probably we can invent a minimal online-QA-import tool - for checking POI imports via uploading</div><div>- house number, street name, postcode, city name: validation</div><div>- Search for nearby Mapillary/* images</div><div>- Check/show different building names vs. candidate poi names inside ( <a href="https://gist.github.com/ImreSamu/68b8cdaabcd7098a88cfa2ee96c47b12">https://gist.github.com/ImreSamu/68b8cdaabcd7098a88cfa2ee96c47b12</a> )</div><div>- Check/show similar names with different tags</div><div>- visual checking on the maps.</div><div>- visualize the osm files as a table .. </div><div>- ...</div><div><br></div><div><br></div><div>And Thank you for your work,</div><div> Imre</div><div><br></div><div><br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">Hiausirg <<a href="mailto:grussausbw@gmx.de">grussausbw@gmx.de</a>> ezt írta (időpont: 2022. ápr. 21., Cs, 10:37):<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">"Data is from 2009"<br>
Sorry, didn't realize that.. Then it is probably better to use the data only as a reference. However: Before proposing this, I checked quite a lot of the POIs (verified on their websites), and almost all were correct. But yeah.<br>
<br>
Replies to other comments (not that they would really matter now anymore):<br>
"OpenStreetMap is not an IT project, it is a put-your-feet-on-the-ground-and-map-stuff project. Is it really a good idea for you to be adding chicken slaughterhouses to<br>
the map based on blindly trusting a government source?"<br>
I don't even know what an "IT project" is. And while your project definition is true, imports are an important part too. Without them, the US map (or the database) would look totally different - and by different, I mean empty. Also, i'm not "blindly trusting" sources. I inspected the data and had nothing to complain about. Nothing offset, nothing unclear, nothing wrong. The only problem with the source could be that a few things are outdated - ... or the source is outdated. Oops.<br>
<br>
"you might be able to draw more motivation from OSM by mapping things yourself, rather than scouring public datasets for stuff that might be suitable to import"<br>
I mostly map things myself. Power lines, pipelines, (rail)roads - and industry. For example, i mapped almost all of the >115kV power transmission grid and pipelines in Georgia (<a href="https://openinframap.org/#6.79/32.574/-83.812/L,O,P,T,W" rel="noreferrer" target="_blank">https://openinframap.org/#6.79/32.574/-83.812/L,O,P,T,W</a>). But as mentioned, also industrial facilities - including slaughterhouses. So this data would also help me with my editing area/intrest.<br>
<br>
"OSM currently has 11 objects in the US with "product=meat". After your import, there would be 4000, meaning that over 99.6% of all "product=meat" objects will have come from your import. There are currently 50 slaughterhosues in OSM in the US, and after your import there will be 750, i.e. over 90% of all slaughterhouses will be from your import."<br>
Isn't this the best reason FOR the import? The less data we have on that subject, the more sense it makes to import it from good sources, or not?<br>
<br>
"I would strongly suggest not to use a source=*"<br>
I think that this might be useful, so that mappers know directly that this object was a part of an import, without having to find the changeset tags first. And 4k objects are now not that much, compared to the 220 million source tags in the DB (and this number is still continously growing). However, i don't have a problem to omit it either.<br>
<br>
"in the addr:housenumber=* there are a lot of strange values"<br>
Intresting! I think I know where it comes from: From Puerto Rico (also US territory). Their adressing system is somehow different than in the rest of the US. But I would not have imported anything there anyway, only in the continental US (so no outlaying islands).<br>
<br>
"*3322 "TEXT35temp" <---- Strange , please verify ! *"<br>
I left that in, because it was a bit unclear if additional tags for these Processing features were needed. It would be removed before uploading. (on the wiki talk page, but i would have explained this here as well)<br>
<br>
Greetings<br>
<br>
_______________________________________________<br>
Imports mailing list<br>
<a href="mailto:Imports@openstreetmap.org" target="_blank">Imports@openstreetmap.org</a><br>
<a href="https://lists.openstreetmap.org/listinfo/imports" rel="noreferrer" target="_blank">https://lists.openstreetmap.org/listinfo/imports</a><br>
</blockquote></div></div>