Community Building and Collaborative Georeferencing using GEOLocate

1
Community Building and Collaborative
Georeferencing using GEOLocate

• Nelson E. Rios
• Henry L. Bart Jr.
• Tulane University Museum of Natural History

2
What is Georeferencing

• As applied to natural history collection data it
  is the process of assigning geographic
  coordinates to a textually described collecting
  event
• Traditional approaches laborious and time
  consuming
• Automated and collaborative processes have proven
  to improve efficiency

3
GEOLocate Overview
Desktop application for automated georeferencing
of natural history collections data
Initial release in 2002
Locality description analysis, coordinate
generation, batch processing, geographic
visualization, data correction and error
determination
4
Overview Locality Visualization Adjustment
Computed coordinates are displayed on digital maps
Manual verification of each record
Drag and drop correction of records
5
OverviewMultiple Result Handling
Caused by duplicate names, multiple names
multiple displacements
Results are ranked and most accurate result is
recorded and used as primary result
All results are recorded and displayed as red
arrows
Working on using specimen data to limit spread of
results
6
OverviewEstimating Error
User-defined maximum extent described as a
polygon that a given locality description can
represent
Recorded as a comma delimited array of vertices
using latitude and longitude
7
Current Lines of Development

• Taxonomic Footprints
• Collaborative georeferencing
• Global expansion
• Multilingual georeferencing with user-defined
  locality grammar

8
Taxonomic Footprint Validation
Uses point occurrence data from distributed
museum databases to validate georeferenced data
Taxa collected for a given locality
Species A
Species B
9
Footprint for specimens collected at Little
Schultz Creek, off Co. Rd. 26 (Schultz Spring
Road), approx. 5 mi N of Centreville Bibb
County White circles indicate results from
automated georeferencing. Black circle indicates
actual collection locality based on GPS. This
sample was conducted using data from UAIC TUMNH
10
Collaborative Georeferencing

• Distributed community effort increases efficiency
• Web based portal used to manage each community
• DiGIR used for data input (alternatives in
  development)
• Similar records from various institutions can be
  flagged and georeferenced at once
• Data returned to individual institutions via
  portal download as a comma delimited file

11
Collaborative Georeferencing
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
Global Georeferencing

• Typically 11,000,000
• Will work with users to improve resolution
  (examples Australia250K Spain200K)
• Advanced features such as waterbody matching
  bridge crossing detection possible but requires
  extensive data compilation (example Spain)

35
Multilingual Georeferencing

• Extensible architecture for adding languages via
  language libraries
• Language libraries are text files that define
  various locality types in a given language
• Current support for
• Spanish
• Basque
• Catalan
• Galician
• May also be used to define custom locality types
  in English

36
(No Transcript)
37
(No Transcript)
38
Future Directions

• Collaboration with foreign participants to
  improve datasets and language libraries
• Cross platform Java client
• More web services integration
• Integration of WFS WMS for mapping
• Alternatives to DiGIR

39
Final Thoughts

• Approximately 20 of DiGIR providers (excluding
  fish collections) tested have problems of
  stability and/or data quality resulting in the
  inability to cache records.
• Typical data quality issues
• Malformed data (common in datelastmodified)
• Catalog numbers missing and/or duplicated
• Resolution requires dedicated infrastructure for
  data providers or alternative means of serving
  data.

40
Acknowledgements

• Djihbrihou Abibou
• Andy Bentley
• Shwetha Belame
• Paul Flemons
• Demin Hu
• Sophie Hung
• John Johanson
• David Draper Munt
• Dave Vieglais
• Ed Wiley
• National Science Foundation