| Abstract Detail
Systematics Section Macklin, James [1], Morris, Paul J. [2], Morris, Robert, A. [3], Wang, Zhimin [3], Kelly, Maureen [1]. Herbarium Networks Part III: A prototype for exchange of botanical specimen data to reduce duplicative effort and improve quality using a ‘filtered push’. The challenge of digitizing the remaining nearly 90 million herbarium specimens held in US institutions by 2020 remains daunting. Capturing these records with efficiency and high quality is paramount. The existence of duplicate specimens can be taken advantage of if institutions could share data. Data entry personnel would save time by not re-entering information, and improve quality by detecting errors or adding further knowledge that would enhance the value of the record.
A network solution, the Filtered Push, implementing map-reduce has the potential to identify duplicates, improve quality, and enhance communication about specimens between curators and researchers. Progress to date includes a prototype network with a client implemented in Specify Workbench, which can detect duplicates within the network, allow curatorial oversight, and provide for rapid inclusion the institution's master Specify database. To discover related information, the capturer would input a collector name and number and be presented with a reduced set of records, which should match the specimen data they are about to digitize. This interface will allow gated access pending human review of sets of ‘filtered’ information gathered from the community before injection into a herbarium database. We are currently testing the ‘push’ of annotations based on authoritative specimens and/or expert opinions to institutions which house a relevant specimen to enhance their record quality. We are also refining analysis tools to detect related records and use authoritative data sets to discover outlying specimens and potential errors.
An interface to the network could also allow a researcher interested in a particular taxon to query a syndication service that would return data and analyses based on additions and/or changes. This could greatly enhance the annotation of specimens as an expert would be constantly aware of the status of the specimens in the network and be able to efficiently contribute their knowledge.
Log in to add this item to your schedule
Related Links:
Filtered Push Wiki
1 - Harvard University, Department of Organismal and Evoloutionary Biology, 22 Divinity Ave., Cambridge, MA, 02138
2 - Harvard University Herbaria & Museum of Comparative Zoology, Organismic and Evolutionary Biology, 22 Divinity Avenue, Cambridge, MA, 02138, USA
3 - University of Massachusetts, Boston, Computer Science, Boston, MA, USA
Keywords:
herbarium
informatics
specimens
network.
Presentation Type: Oral Paper:Papers for BSA Sections
Session: 72
Location: Wasatch B/Cliff Lodge - Level C
Date: Wednesday, July 29th, 2009
Time: 3:30 PM
Number: 72001
Abstract ID:795 |
