Paul Norman wrote:
I happened across an import of Fresno castradal data from mid-2010 in the
Fresno area. http://www.openstreetmap.org/?lat=36.77&lon=-119.81&zoom=15 is
the general area but I haven't fully explored the extents. For a view of the
data, see http://maps.paulnorman.ca/imports/review/fresno.png
Based on source tags this import was of about 280k ways and their associated
nodes. This likely totals about 1M objects.
The tagging of a typical lot is as follows:
attribution=Fresno_County_GIS
fresno_APN=44329204
fresno_TX_area_CD=5199
fresno_lot_area=0.17
fresno_lot_depth=124.0
fresno_lot_width=60.0
landuse=residential
lot_description=LOT 13 CLINTON TERRACE NO 2
lot_type=single family residential properties
other_use=S
primary_use=S01
secondary_use=000
source=http://www.co.fresno.ca.us/%5Cdepartmentpage.aspx?id=16313
Many lots have the fresno_lot* tags as 0.0. This is clearly absurd and the
object's tags are inconsistent with its geodata.
There are a number of problems with this data. These include
1. It is castradal data. The consensus is against dumping castradal data
into OSM.
2. The tagging hasn't been converted to OSM values. The other_use,
primary_use and secondary_use should map to something or perhaps not have
been imported.
3. There are tags indicating the size of the lot. This duplicates the OSM
geodata
4. Some lots are split in half, for example the school in the picture.
5. There aren't many curves, but what curves I can find are overnoded.
6. Some objects have no OSM tags at all. For example,
http://www.openstreetmap.org/browse/relation/957260 or
http://www.openstreetmap.org/browse/way/61615684 have no tags except for
metadata from the import.
7. Some objects don't appear to correspond to anything on the ground.
http://maps.paulnorman.ca/imports/review/fresno2.png is an example of this.
The import is approximately the same age as the Bing imagery.
8. There are duplicate nodes where data was imported on top of other data.
For example, http://www.openstreetmap.org/browse/node/768314177
http://www.openstreetmap.org/browse/node/767799968
http://www.openstreetmap.org/browse/node/767770150
With all of these problems I cannot think of any ways to fix the problems
short of reverting the import. The tagging problems could be fixed by a
script but the inherent problems of castradal data cannot be fixed without
essentially deleting most of the import anyways.
I propose to delete unmodified objects from this import. I will attempt to
preserve areas like schools and fix them if possible. It should be possible
to keep most of them but I won't be able to be sure until I get into the
removal.
Such a removal would have to be timed around the import and mechanical edit
restrictions that will be in place during the rebuild process. If it
occurred before or after the rebuild would depend on when the rebuild
process starts and how long the consultation about this proposal takes.
_______________________________________________
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us
User nmixter has been the user who did the import. I would recommend to
revert the changeset(s) and delete the useless stuff. In the small area
I checked there were many errors (overlapping lines, double nodes...). I
agree, that there is no way to fix stuff. User BiIbo modified many
objects (about 33 %), but it is not obvious what he really changed.
There are amenity=place_of_worship with religion=xyz or amenity=. What
do you recommend ? We could keep amenity=school or place_of_worship and
delete the rest of the tags. There are amenity=public_facility. There
exists an amenity=public_building, but in many cases public_facility is
used not only for buildings.
We also have landuse=farm or landuse=wood. Sometimes we have
landuse=farm combined with natural=water.
I think we should simply delete all objects without any osm-tags.
Tell me if you need help.
WernerP
_______________________________________________
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us
