[OSM-talk] Element on OSM which don't exist in real life

[Sebastian Arcus]

On 02/12/13 20:26, Richard Welty wrote:
>> Thank Frederik. I'm actually working in JOSM, and I've spotted those
>> import tags - but as I still feel like a newbie when it comes to all
>> things OSM, I thought I'd check with the community so that I don't do
>> something silly.
> Welcome to TIGER and GNIS!
>
> what Fredrick said is true; i'm going to add some US specific info.
>
> there is some bad data in the GNIS import. when i encounter obviously
> bogus objects from the GNIS import, i just delete them. just the other
> day i deleted a GNIS object which suggested someone had a heliport in
> their back yard a little south of Albany NY. i didn't see a heliport.
>
> as for non-existent roads in TIGER, it happens. there are various
> reasons for it, but if you're looking at unreviewed TIGER (look for
> a tiger:review=no tag) and you can't find the matching real world
> entities, you are fully justified in deleting them. if you do verify
> a road exists, then also verify its topology (is it hooked up the way
> the map shows), its name (from the road signs) and its location (bing
> imagery is good for this, as are personally collected GPS tracks).
> if those check out you are justified in removing the tiger:reviewed
> tag (if you are fastidious, you can change it to yes, but most of us
> just delete it. fewer mouse clicks that way.)
>
> as for the reasons - TIGER data quality varies quite widely, seemingly
> from county to county. it is getting better - if you compare the 2005
> data which is what we imported to 2013 data you can see the improvement.
> but there are cases where obviously someone sketched out a map then
> freehanded it into the database, and there are cases where a developer
> got their roads into the database before they failed to finance the
> development project.
>
> if you look at the image dropdown, you will see a tiger 2012 underlayment.
> you can use this to compare what's in OSM to the 2012 data set.
>
> so don't be afraid to fix it. there is bad data in TIGER. mistakes
> happen.
>
> richard
>
Thanks Richard. Very helpful explanation and pointers.

Sebastian