GEOLocate Georeferencing Software for Natural History Collections

Natural Language Processing

The algorithm used by GEOLocate begins by standardizing the locality string into common terms and parsing out distances, compass directions, and key geographic identifiers. This information is then used in a series of lookups and displacement calculations to determine geographic coordinates. Placename, river mile, legal land description and higway-waterbody crossing datasets are used for lookups. Displacements from these lookups are calculated if indicated by the parsed locality information. Coordinates output from the initial georeferencing may be further refined via an additional function to scan the locality string for waterbody names and “snap” output coordinates to the nearest point on the waterbody found. This feature has proven very useful for aquatic collections. The resulting coordinates are then ranked based on the type of information found within the string and plotted on the digital map display for user verification, correction and error determination. You can try out an online version of the process here or request a free copy of the full featured standalone version here.

The following are an actual sample of data georeferenced by GEOLocate:

Arkansas River at River Mile 10 boat landing., USA, Arkansas, Desha
Green River at Roachville ford approximately 2 mi. E. of Greensburg, USA, Kentucky, Green
Alabama River at Wilcox Bar; River Mile 120., USA, Alabama, Wilcox
Missouri River 3 mi. SE of Pierre, South Dakota., USA, South Dakota, Hughes
Natalbany River at U.S. Hwy. 190, USA , Louisiana, Tangipahoa
Tussahaw Creek at LeGuin Mill Road, approximately 3.5 mi. ENE Locust Grove - Segment 3., USA, Georgia, Henry
Escatawpa River at Hwy. 612., USA, Mississippi, George
South Fork Little Red River at Arkansas Hwy. 95, SW Clinton, Section 11., USA, Arkansas, Van Buren
Little Pine Barren Creek at Hwy. 99; T4N R32W Sec. 4., USA, Florida, Escambia
t1N r3e sec. 13, USA, Nebraska, Jefferson

User Interface

One of our goals was to provide an interface by which users could georeference records one by one or in batches from files, vizualize and correct calculated coordinates and determine polygonal error descriptions. GEOLocate uses XML as its native file format but also supports data import from .CSV and delimited .TXT files. Once coordinates have been derived from a locality description adjustments may be made by simply click and dragging a displayed point on a map. Error estimates can then be recorded as the maximum extent which a description could occupy. This extent is represented as a comma delimited array of polygon vertices and can easily be drawn onto the map.