Symbiota (http://symbiota.org/tiki/tiki-index.php) is an open source software designed to promote and facilitate
collaboration among those working to document biodiversity. Symbiota has become increasingly popular in
recent years in North America, due in part to its suitability to support large herbarium networks and
NSF-sponsored Thematic Collections Networks (TCNs; see https://www.idigbio.org/content/thematic-collectionsnetworks).
The specimen-based Content Management System (CMS) provides a shared platform allowing
researchers to manage biological resources as an integrated network. Data management through a
community-based system has allowed for the development of several features and workflows that have
enhanced efficient data entry while improving overall data integrity and quality. On-line data entry directly from
an image of the specimen label allows for label transcription and error resolution that can call upon a global user
community. A novel crowd sourcing feature in Symbiota offers collection managers the ability to submit
specimen label images to a queue for group data entry by a volunteer task force. To improve efficiency and
quality, the user interface incorporates Optical Character Recognition (OCR) and Natural Language Processing
(NLP) capabilities, as well as duplicate and exsiccati record harvesting and real-time data validation. The
duplicate clustering module groups duplicate specimen records across institutions, thereby obviating the need to
re-enter a previously processed specimen and enhancing the task of locating and resolving misidentified
specimens, viz. by highlighting the most recent annotation events within a cluster. As an additional review step,
collections can opt to allow registered users to fix basic errors if and when they encountered them. Collection
managers have the ability to review, approve, or revert such edits. Several other novel community features are
available through Symbiota, including an integrated loan management module and pre-accessioned data entry
by the original collector. We will demonstrate and discuss these features, their underlying concepts,
implementation, utility, and future steps to further augment the community of contributing users.
Franz Et Al. Crowd Sourcing and Community Management Capabilities Available within Symbiota Data Portals
1. Crowd Sourcing and Community Management
Capabilities Available within Symbiota Data Portals
Nico Franz1, Corinna Gries2 , Thomas Nash III2 & Edward Gilbert1
1
School of Life Sciences, Arizona State University
2 Center for Limnology, University of Wisconsin
TDWD 2013 Annual Conference, Florence, Italy
Building and Maintaining Crowd Sourcing Websites and Their Communities
October 29, 2013
Presentation Overview @ http://taxonbytes.org/tdwg-2013-crowd-sourcing-and-community-management-capabilities-with-symbiota/
2. Lichen, Bryophytes and Climate Change – scope and goals
• Support – NSF ADBC Program – Award EF 1115116
• Covering ~ 2.3 million specimens: 900,000 lichens & 1.4 million bryophytes
• 90% of all specimens housed in this region; > 60 non-governmental herbaria
• LBCC has 16 focused digitization centers where voucher labels are imaged
= LBCC imaging centers
3. LBCC sustaining Symbiota-based data portals
• Consortium of North American Lichen Herbaria
• URL: http://lichenportal.org/portal/
• Currently with 51 member collections
• Total of 1,084,888 records (October, 2013)
CNALH
• Consortium of North American Bryophyte Herbaria
• URL: http://bryophyteportal/
• Currently with 46 member collections
• Total of 1,437,735 records (October, 2013)
• Each portal is sustained by Symbiota
CNABH
4. LBCC member portals are active virtual environments
Lichen portal – 7,302 visitors / 3 months
Bryophyte portal – 1,530 visitors / 3 months
5. Overview of the LBCC digitization workflow (label imaging)
• The LBCC digitization workflow (completion of records) depends critically on the
sustained participation of editors/transcribers, and of volunteers.
6. Image URLs
Herbarium
Database
Capture Image
barcode in file
name
Create Skeleton
File
species
name, country,
state, exsiccati,
etc.
Imaging Stage
Detailed workflow diagram
Workflow "closes" in home collection
Upload to
FTP server
Upload to
FTP server
Create New Record
barcode, image, ske
letal data
Symbiota
Editor
review, edit, ke
ystroke
Manage
Specimen Data in
Portal
Image processing
extract
barcode, create
web versions, map
to portal DBs
Existing Record
simply link
image
Manage / Review
Records in Portal
Automated OCR
Tesseract, ABBYY
Crowdsourcing
Central
LBCC Crowdsourcing elements
Automated NLP
Darwin Core Parsing
11. "Transcribe the ALCAN Expedition"
"Create your own…" *
* Instant feedback on data volume.
12. Listing of pending "Harriman" records;
each Symbiota ID is clickable to edit.
13. LBCC
Crowd
Sourcing
Central
Record
1545184
• The LBCC digitization workflow pipeline has produced
a "skeletal record", including:
• Record GUID
• Thesaurus-ratified Scientific Name (not editable)
• OCR of voucher locality label image
• "Parse OCR (LBCC)" [a custom LBCC program] will get
the transcription process underway.
15. 2. Correction of the parse using Symbiota tools (e.g. GeoLocate)
16. 3. Approaching a clean record1 transcription, ready for saving
1
DwC Class: CleanRecord – Utter these two words in front of a TDWG
audience, then immediately prepare to… [remainder not yet ratified].
17. Crowd Sourcing Central – Score Board *
Options to review one's submitted
records and review points assigned
(by the collection's manager).
* See also Appendix I.
18. Crowd Sourcing Central – User's Review Pages *
My pending records with LBCC.
My 2/4 approved records/points.
* See also Appendix III.
19. Crowd Sourcing Central – Collection Manager's Control Panel
4215 newly digitized records are
available for addition to the queue.
25 submissions pending.
* See also Appendix II.
20. Crowd Sourcing Central – Collection Manager's Review Pages
2 points = default score.
Specific feedback possible.
* See also Appendix III.
21. Lichen, Bryophytes and Climate Change – CS in review
• A key purpose of the LBCC portal CS entry environment is to create a user
experience that is personalized.
• Special expeditions are a subset of the records queue for CS data entry, and are
identified as being part of a "special group/theme" of specimens.
• Expeditions are meant to educate those who are performing the data entry
about a specific event.
• They also aid data entry because the user generally deals with a homogeneous
type of label format, as opposed to shifting between numerous layout types.
• User input and managerial control (review, feedback, scoring) are interactively
facilitated in the same Crowd Sourcing module implemented in Symbiota.
22. Acknowledgments
• TDWG 2013 Symposium organizers – Paul Kenneth Flemons
• Ben Brandt & John Brinda – LBCC software development
• Participating CNALH & CNABH collections
• NSF Award EF-1115116.
"Digitization TCN – Collaborative Research: North American Lichens and
Bryophytes: Sensitive Indicators of Environmental Quality and Change."
http://lbcc1.acis.ufl.edu/
http://lichenportal.org/portal/
http://bryophyteportal/
https://sols.asu.edu
http://symbiota.org/tiki/tiki-index.php
http://taxonbytes.org
23. Appendix I: Crowd Sourcing Central "rules of engagement"
1. Available at http://lichenportal.org/portal/collections/editor/crowdsource/central.php
2. Shows scores and collections participating in crowd sourcing, along with their statistics.
3. Available to all viewers of the site, irrespective of whether they are logged in.
4. If users log in, then their scores will be displayed in a separated information "box".
5. The link above will generally – on most portal sites – be added to the main left menu, or
made available from another crowd sourcing page that is custom generated for a
project. For instance, LBCC will likely link to this page from their main Drupal page.
6. Clicking on "review records" within the Current User's Standing box will take the user to
the Review page (see Appendix III).
7. Clicking on numbers within the collection table will take the user to a list of specimens
queued up for data entry (and open specimens within the CS queue).
24. Appendix II: Collection Manager's Crowdsourcing Control Panel
1. Available at http://lichenportal.org/portal/collections/editor/crowdsource/controlpanel.php?collid=22
2. Available only to collection managers.
3. Shows statistics only for a given collection.
4. Available also from the collections control panel (not yet implemented in the public
site) and in Crowd Sourcing Central (via the editing symbol to the right of the collection
names).
5. Allows managers to edit crowd sourcing instructions or link to a training URL.
6. A link to the right of "Available to Add" is where a collection manager would add their
records to the Crowdsourcing queue.
25. Appendix III: Review Pages for contributors, managers
1. Available from a collection manager's perspective or a user's perspective, yet behaves
somewhat differently depending on the perspective.
2. Collection manager perspective:
1.
Main purpose is to enable a quick review of specimen records that are pending (or re-review of closed
records).
2.
Available from the Collection Manager’s Crowdsourcing Control Panel by clicking on the "Review" link to
the right of the numbers.
3.
A collection manager can assign points to an annotated record (2 points is the default
value), comment, and change the CS status to closed (approved).
4.
Managers can edit all records, whether they are pending or closed.
3. User perspective:
1.
Available from Crowd Sourcing Central by clicking on "Review Records" within the Current standing box.
2.
Allows user to review and access records with pending status.
3.
Allows user to review points and provide comments for closed records.
4.
Users can edit all pending records.
5.
Users can review yet not edit records that have been closed by a collection manager.