I ran a frequency on the original shapefile and started cleaning up
misspellings for the OSM conversion (the shapefile I provided was before
I started doing any corrections). I contacted Union County and alerted
them to the frequent spellings of commercial/residential and a few other
assorted oddities. I pulled churches and schools out of the original -
those needed an extra tag: amenity = place_of_worship and I think
amenity = school. I mean to have placed those up there and forgot
completely.
I plan on fixing any errors with buildings crosses buildings on a case
by case basis.
I ended up manually expanding a lot of the street tags - there are a few
I'm still not clear on. I've got an email into them to check.
One of the things I'm puzzled by (just from a mapping perspective) are
the houses with an address and those without. It would appear they are
still in the process of assigning addresses to their data. They provided
a 911 address file but it left me with more questions than answers -
I'll zip them an email to make doubly sure the address on the building
is correct. Some of their road data crosses county/state lines and I
think that ended up being incorrectly pushed into the building footprint
(i.e. the road originates in fanin county or in North Carolina).
I don't know how multiple addresses are handled but I will check with
them and get an answer.
I appreciate the help. I'm still a while from wanting to start an upload
- I want to clean up the base data before this starts.
Randy
On 12/08/2013 07:23 PM, Jason Remillard wrote:
Hi Randal,
Some comments.
- The alignment matches Bing very closely.
- There are about 100 building inside building, and 2 crossing
buildings errors from the JOSM validator. It looks like the error is
in the source data.
- What is your mapping to the building tag? The source data has all
kinds of misspellings. 7 spellings for residential. It also has fields
for mobile homes, school, church, government, that have specific
building tags that are not in the output file.
- You have some weird house numbers in the source data that need to be
filtered out (0, "burned down", not addressed, not livable, etc).
- What is your mapping for the street_T field. There are some weird
things in the data like '<Null>', 'DR (FANNIN)', 'PL','RD W", 'RDG',
etc that are making it through to the output file.
- Same for the street. There is some weird values, like '<null>', "12
point", etc.
- Dr is not being expanded to Drive.
Do you know how multiple addresses at the same building are handled
in the source data?
Thanks
Jason.
On Sun, Dec 8, 2013 at 3:20 PM, Randal Hale
<[email protected]> wrote:
I placed two links here:
http://wiki.openstreetmap.org/wiki/Data_Import_for_Union_County_Georgia
One is for the original shapefile data and one is for the OSM file.
The original shapefile is in ESRI:102667
(http://spatialreference.org/ref/esri/102667/) - which is equivalent to
EPSG:2240
I converted the entire file into OSM format. It should be in EPSG:4326
I've started cleaning up the tiger/NHD/landcover data currently in Union.
I'm using the roads file that was provided for guidance. The area is
exceptionally rural. I'm slowly removing landcover since it's almost
entirely wrong. NHD is being fixed where it makes sense - being removed in
cases where it no longer matches the imagery. I'm using 2010 NAIP as a
reference in the cleanup.
--
-----------------
Randal Hale, GISP
North River Geographic Systems, Inc
http://www.northrivergeographic.com
423.653.3611 [email protected]
<mailto:[email protected]>
twitter:rjhale
http://about.me/rjhale
_______________________________________________
Imports-us mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/imports-us
--
-----------------
Randal Hale, GISP
North River Geographic Systems, Inc
http://www.northrivergeographic.com
423.653.3611 [email protected]
<mailto:[email protected]>
twitter:rjhale
http://about.me/rjhale
_______________________________________________
Imports-us mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/imports-us
