Frequent data errors for negative (western) longitude geo locations on de.dbpedia.org

Data Data Quality
1

Hi,
i just discovered data errors for negative (western) longitude geo locations on de.dbpedia.org.
western longitude is indicated by negative numbers however the negative sign seems to be missing on many dbpedia entries for de. Values are correct for dbpedia.org

for example:
(sorry had to change link host names because direct linking not allowed in post)

de_dbpedia_org /page/Los_Angeles (incorrect)
georss:point
34.05222222222222 118.0
dbpedia_org/ page/Los_Angeles (correct)
georss:point
34.05 -118.25

de_dbpedia_org /page/San_Francisco (incorrect)
geo:lat
37.779301 (xsd:float)
geo:long
122.000000 (xsd:float)
dbpedia_org /page/San_Francisco (correct)
geo:lat
37.783333 (xsd:float)
geo:long
-122.416664 (xsd:float)

de_dbpedia_org /page/Hawaii (incorrect)
21.311388
157.000000

dbpedia_org /page/Hawaii (correct)
21.311388
-157.796387

de_dbpedia_org /page/Idaho (incorrect)
44.416668
114.000000

This looks like a systematic problem, however there are also correct entries:
de_dbpedia_org /page/Alaska
georss:point
65.0 -151.0

These are just some random examples i came across, guess it should be possible to find out the real scale of this problem with a sparql query?

2

@christian,
thanks for reporting, the error seems to persist in the October release (see here https://databus.dbpedia.org/dbpedia/mappings/geo-coordinates-mappingbased/2020.10.01 or here https://release-dashboard.dbpedia.org/?version=2020.10.01#version)

 curl https://downloads.dbpedia.org/repo/dbpedia/mappings/geo-coordinates-mappingbased/2020.10.01/geo-coordinates-mappingbased_lang=de.ttl.bz2 | bzcat | grep '/San_Francisco>'
<http://de.dbpedia.org/resource/San_Francisco> <http://www.georss.org/georss/point> "37.7793 122.0" .
<http://de.dbpedia.org/resource/San_Francisco> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2003/01/geo/wgs84_pos#SpatialThing> .
<http://de.dbpedia.org/resource/San_Francisco> <http://www.w3.org/2003/01/geo/wgs84_pos#lat> "37.7793"^^<http://www.w3.org/2001/XMLSchema#float> .
<http://de.dbpedia.org/resource/San_Francisco> <http://www.w3.org/2003/01/geo/wgs84_pos#long> "122.0"^^<http://www.w3.org/2001/XMLSchema#float> .

@fab_hop, @m1ci: could we test the new debugging workflow for this?

3

Hi,
thanks for your reply.

Here is the list of errors i found so far, hope that helps:

Los Angeles
San Francisco
Kanada
Taos Pueblo
Mesa-Verde-Nationalpark
Utah
Salt Lake City
Carmel-by-the-Sea
Rocky Mountains
Mexiko
Oʻahu
Hawaii
Idaho
Glacier-Bay-Nationalpark

also invalid but probably not related to negative longitude error:
Cilento
Schwäbische Alb

i found another issue related to location:
http://de.dbpedia.org/page/Neuseeland
geo:lat
-47.290279 (xsd:float)
-46.677776 (xsd:float)
-40.843613 (xsd:float)
-34.396389 (xsd:float)
52.000000 (xsd:float)
geo:long
167.000000 (xsd:float)
169.000000 (xsd:float)
172.000000 (xsd:float)
173.000000 (xsd:float)

the 5th latitude attribute is clearly wrong.
when parsing dbpedia locations, do i have to expect this kind of box coordinates? seems like most of the other locations are center based single-spot.

4

Hi @kurzum
is there any chance this will get fixed in the near future?
just checked Los Angeles, Mexiko and Kanada - still wrong longitude attributes…

thanks,
christian