a platform for georeferencing natural history collection data

A Platform for Georeferencing Natural History Collection Data

Nelson E. RiosYale University Peabody Museum

What is Georeferencingpushepatapa creek, trib. to pearl river, 7.8

miles north of bogalusa at hwy 21; Washington; LA; USA

latitude: 30.88797longitude: -89.83601

uncertainty radius: 48muncertainty polygon: 30.88823,-89.83641, 30.88815,-89.83634,

30.88808,-89.83622…

The Problem

3 billion specimens with ½ billion collecting events

Software & services for georeferencing of natural history collections data

automated georeferencing

verification & correction

batch processinggeographic visualization

uncertainty determinationcollaborative georeferencing

interoperability

multi-lingual

kml export

google, bing, openstreet, wms

soap & rest api

training

Notable Milestones• 2003 Initial release as desktop application• 2006 Global scope• 2008 Multilingual capabilities, CoGe platform• 2009 Web Services & third-party integration • 2010 Web Clients & third-party integration• 2012-2015 FishNet - CoGe improvements• 2016 SERNEC - Symbiota CoGe communication

protocol• 2017 Historical mapping• 2018 Move to geo-locate.org• 2019 Integration with EMu• 20XX Decentralized Data Annotation??

Typical GEOLocate W

orkflow

pushepatapa creek, trib. to pearl river, 7.8 miles north of bogalusa at hwy 21; Washington; LA; USA

Georeferencing Algorithm

Visualize, verify & adjust output coordinates & uncertainties

Data Entry &Preparation

AutomatedProcessing

ManualVerification

latitude: 30.88797longitude: -89.83601

uncertainty radius: 48m

uncertainty polygon: 30.88823,-89.83641, 30.88815,-89.83634, 30.88808,-89.83622…

1

2

3

Processes

Collecting Event Textual Data

Candidate Spatial Representations

Accepted Spatial Data

Collection Management System

Collection Management System

Verification and refinement

NLP & Geospatial Analysis

Repatriation

Data preparation

(Time vs Quality)

Interactions

Web ServicesGeospatial Descriptors (Latitude, Longitude, Uncertainty, Polygons etc)

Basic Client Batch Client Collaborative Client

Collaborative Data Portal

Embeddable Client

Third Party Clients (Emu)

Third Party Clients CMD Line Client Clients

Third Party Clients (Symbiota)

Locality Records

Specimen Records

Local Interfaces

Named Place

7 miles N

orth of

Named Place

7 miles N

orth of

Creek at Hwy 1

Match Water Body Detect Hwy/River Crossing

Point

Point & Radius

Point, Radius & Polygon

Moving Beyond Points

Generating Polygons:

15 mi N of Lawrence, KSdisplaced polygon

15 mi N of Lawrence, KSno displacementLawrence, KS

Geospatial Representation - Polygons

POLYGON ((-94.480762 36.174037, -94.451064 36.205624, -94.427375 36.172374, -94.480762 36.174037))

36.174037,-94.480762,36.205624,-94.451064,36.172374,-94.427375,36.174037,-94.480762

36.174037-94.480762

36.205624-94.451064

36.172374-94.427375

GLC Format

WKT Format

Visualization: Base Layers

Application Services: Web Client APIs

• URL API for user input & lightweight web clients

• Lightweight client, specifically designed for embedding into other web applications.

• Two way communication between web sites uses JavaScript postMessage() • Compatible with all modern browsers:

IE 8.0+Firefox 3.0+Safari 4.0+Chrome 1.0+Opera 9.5+

EMu (desktop)Specify (web)

Symbiota (web)Arctos (web)

Speeding things up

• Specimen records• Collecting events• Grouped collecting events• Data prioritization• Locality services pre-processing• Expert georeferencers• High-throughput (first pass georeferencing)

• Processes for automated QA/QC• Collaborative georeferencing

locality required the locality description country required state optional will return zero results if country is USA and state is not provided county optional hwyX optional enable looking for intersections between roads and water bodies. values: [true], false enableH2O optional snap points to water bodies if found in locality.

values: [true], false doUncert optional calculate uncertainty radius.

values: [true], false doPoly optional generate polygons for known places.

values: true, [false] displacePoly optional displaces polygons in line with locality description. doPoly must be enabled, otherwise ignored.

values: true, [false] languageKey optional selected language. 0=English.

values: [0],1,2,3,4 fmt optional output format.

values: [json], geojson callback optional wraps response in a callback

Examples: http://www.geo-locate.org/webservices/geolocatesvcv2/glcwrap.aspx?Country= USA&Locality=bogalusa&state=la http://www.geo-locate.org/webservices/geolocatesvcv2/glcwrap.aspx?Country=USA&Locality=bogalusa&state=la&dopoly=true&fmt=geojson

Comments: Currently errors are returned as plain text without any JSON formatting. This will be updated in a future update. For example (when not enough data provided): http://www.geo-locate.org/webservices/geolocatesvcv2/glcwrap.aspx?Country=USA

url: http://www.geo-locate.org/webservices/geolocatesvcv2/glcwrap.aspx?

RESTful Web Services

High-throughput (“first-pass” georeferencing)• Access geolocate webservices to batch process large datasets

Interoperability & Collaborative

Georeferencing

Collaborative Georeferencing Real-Time Repatriation