How Beneficial are    Data Imports to OpenStreetMap? –       A United States Case StudyDennis Zielstra1 , Pascal Neis2, Ha...
Motivation   Several studies within recent years about OSM data quality in    comparison to governmental or proprietary d...
Main Goals   Evaluate development of OSM data in comparison to    TIGER/Line Data   Determine active/inactive areas in t...
Information about TIGER/Line Import to OSM   First import attempt (2005/06)    TIGER/Line 2005 data (ASCII format)    D...
Data Pre-Processing                      5
General US Data and Member Development                                         6
General Data and Member Development  Members who created at least one Node: 203000 (June 2012)  Members who created most...
General Data and Member DevelopmentMembers with their first node in the US (October 2012)                                 ...
General US Data and Member Development   Strong decrease in residential class due to retagging and data    corrections   ...
General Data and Member Development   Strong improvement in pedestrian and cyclist related data                          ...
Car Data Development per State   Total length comparison of car navigation related information    between TIGER/Line (200...
Car Data Development per Urban Area                        • All urban areas show negative absolute and                   ...
Pedestrian Data Development per State   Total length comparison of pedestrian navigation related    information between T...
Pedestrian Data Development per Urban Area                       • All urban areas show positive absolute and             ...
Geometry Errors and Positional Inaccuracy    Additional problems with imported data:        TIGER/Line 2005 data is at t...
Summary  TIGER/Line import was successfully accomplished during   second attempt, but…      2005 TIGER/Line data importe...
Future Work   Investigate in more detail which data types (geometries,    attributes, feature classes) have been collecte...
Discussion   Does data import discourage mappers from improving car    related road network?   JOSM and Potlatch support...
Thank you   Related papers for the US (http://mygeomatics.com/people/dennis-zielstra/) Zielstra, D., and Hochmair, H. H. ...
Nächste SlideShare
Wird geladen in …5
×

How Beneficial are Data Imports to OpenStreetMap? – A United States Case Study

3.142 Aufrufe

Veröffentlicht am

0 Kommentare
1 Gefällt mir
Statistik
Notizen
  • Als Erste(r) kommentieren

Keine Downloads
Aufrufe
Aufrufe insgesamt
3.142
Auf SlideShare
0
Aus Einbettungen
0
Anzahl an Einbettungen
5
Aktionen
Geteilt
0
Downloads
0
Kommentare
0
Gefällt mir
1
Einbettungen 0
Keine Einbettungen

Keine Notizen für die Folie

How Beneficial are Data Imports to OpenStreetMap? – A United States Case Study

  1. 1. How Beneficial are Data Imports to OpenStreetMap? – A United States Case StudyDennis Zielstra1 , Pascal Neis2, Hartwig H. Hochmair1Geomatics Program – University of Florida1Geoinformatics Research Group - University of Heidelberg2 State of The Map USA , Portland (OR), October 13-14, 2012
  2. 2. Motivation  Several studies within recent years about OSM data quality in comparison to governmental or proprietary data.  Mostly with focus on Europe (i.e. UK, Germany) with no data imports.  Usually good results in countries with active OSM community.  Very little research about OSM US and its data import. 2
  3. 3. Main Goals  Evaluate development of OSM data in comparison to TIGER/Line Data  Determine active/inactive areas in the US by analyzing 50 States (+ D.C.) and 70 urban areas (population > 500000)  Analyze the data source and feature type that was actively contributed 3
  4. 4. Information about TIGER/Line Import to OSM  First import attempt (2005/06)  TIGER/Line 2005 data (ASCII format)  Data was purged in November 2006 due to data integrity problems  Second import attempt (2007/08)  TIGER/Line 2005 data (ASCII format)  “It is unlikely that the TIGER data ever will be imported again.” (Discussions about updated imports in untouched areas)  TIGER 2007 municipal and CDP (census-designated place) boundaries imported  TIGER 2010: ”The quality of this data has improved considerably from the TIGER 2005 data that was originally imported into OSM.” (TIGER 2010 OSM Wiki) 4
  5. 5. Data Pre-Processing 5
  6. 6. General US Data and Member Development 6
  7. 7. General Data and Member Development  Members who created at least one Node: 203000 (June 2012)  Members who created most of their Nodes in the US: 18500 (June 2012) 7
  8. 8. General Data and Member DevelopmentMembers with their first node in the US (October 2012) 8
  9. 9. General US Data and Member Development  Strong decrease in residential class due to retagging and data corrections 9
  10. 10. General Data and Member Development  Strong improvement in pedestrian and cyclist related data 10
  11. 11. Car Data Development per State  Total length comparison of car navigation related information between TIGER/Line (2008-2012) and OSM (2008-Sep. 2012) • Almost all states show negative absolute and relative difference values for OSM (2008 through 2012) => OSM has worse coverage • 7 states with negative difference > 30% in 2012 (e.g. West Virginia and Ohio) • Large negative difference increase in 2010 (average 18%) Example: West Virginia OSM 2012 TIGER 2012 11
  12. 12. Car Data Development per Urban Area • All urban areas show negative absolute and relative values for OSM 2012 vs. TIGER 2011 => OSM has worse coverage • Only 2 urban areas with difference > 30% (i.e. Jacksonville and Salt Lake City) • large difference increase in 2010 (average 16%) Example: Chicago Suburbs OSM 2012 TIGER 2012 12
  13. 13. Pedestrian Data Development per State  Total length comparison of pedestrian navigation related information between TIGER/Line (2008-2012) and OSM (2008- Sep. 2012) • Almost all states show positive absolute and relative difference values for OSM (2012) => OSM has better coverage • 28 states with absolute difference > 1000km (California up to 16600 km) • relative difference as high as 97% Example: California (Yosemite Park) OSM 2012 TIGER 2012 13
  14. 14. Pedestrian Data Development per Urban Area • All urban areas show positive absolute and relative values for OSM 2012 vs. TIGER 2011 => OSM has better coverage • 22 of 70 urban areas with absolute difference > 300 km (Washington, D.C. 1815 km) • relative difference sometimes above 100% Example: Washington D.C. OSM 2012 TIGER 2012 14
  15. 15. Geometry Errors and Positional Inaccuracy  Additional problems with imported data:  TIGER/Line 2005 data is at times inaccurate  Positional inaccuracy not due to shift (appear to be random)  Some areas show geometry errors (blue circle) Example: West Virginia OSM 2012 TIGER 2012 15
  16. 16. Summary  TIGER/Line import was successfully accomplished during second attempt, but…  2005 TIGER/Line data imported in 2007 = outdated.  Results showed severe increase in missing data coverage for car navigation related data in OSM in 2010.  Outdated 2005 TIGER/Line data causes geometry and positional accuracy errors in OSM dataset.  However…  Active data contributions are improving the dataset every day.  Especially pedestrian and cyclist related information is being collected and/or attributed. 16
  17. 17. Future Work  Investigate in more detail which data types (geometries, attributes, feature classes) have been collected and imported. Other imported datasets?  Similar pattern to European OSM quality? (urban areas more active = better quality) 17
  18. 18. Discussion  Does data import discourage mappers from improving car related road network?  JOSM and Potlatch support tiles with TIGER 2011data to trace the new TIGER data. Is this enough?  Discussion about reimporting untouched areas. What about higher quality and more recent data than TIGER in certain states?  What does the OSM community think? Did the imports help or make work more cumbersome? Impact on motivation?  Should there be more restrictions or limitations on imports? 18
  19. 19. Thank you Related papers for the US (http://mygeomatics.com/people/dennis-zielstra/) Zielstra, D., and Hochmair, H. H. (forthcoming). Comparison of Shortest Path Lengths for Pedestrian Routing in Street Networks Using Free and Proprietary Data. Transportation Research Board: Journal of the Transportation Research Board. Zielstra, D. and Hochmair, H. H. (2011). A Comparative Study of Pedestrian Accessibility to Transit Stations Using Free and Proprietary Network Data. Transportation Research Board: Journal of the Transportation Research Board, Volume 2117, pp. 145-152. OSM Weekly News @ http://opengeodata.org/  @den_uf dzielstra@ufl.edu  @pascal_n neis@uni-heidelberg.de Feel free to contact us! 19

×