Semantically-enabled facetted search

One of the limiting factors for today's collaborative science approaches is the difficulty in finding related data sets authored or collected by other groups. The WHOI-hosted NSF Biological Chemical Oceanographic Data Management Office (BCO-DMO) is developing a smart-search capability to meet some of this need. The WHOI Informatics Working Group and RPI staff are working closely with BCO-DMO to develop an easier way for scientists to discover and access data sets related to their research. A part of this collaboration is the investigation into generalizing this framework for use by other WHOI research areas and investigators.

The Biological Chemical Oceanography Data Management Office (BCO-DMO) was created to store and provide access to data sets for marine biogeochemical, ecological and oceanographic data, and information developed in the course of scientific research in support of researchers funded by the National Science Foundation's Biological and Chemical Oceanography division. Data in BCO-DMO can easily be disseminated, protected, and stored on short and intermediate time-frames, and they support the scientific community through improved access to ocean science data. This project is a follow-on project to the successful US GLOBal Ocean ECosystems Dynamics (US GLOBEC) experiment and the US Joint Global Ocean Flux Study (US JGOFS).

One of the first collaborative opportunities identified by the Ocean Informatics Working Group was to explore the use of semantic technologies to support faceted search as part of the BCO-DMO software suite to extend its search and dissemination capabilities, thereby extending the possible science use of the stored data. Faceted searches allow lookups to be performed using a set of classifications that can be ordered in multiple ways. For example, data could be searched by geographic area, then by instrument type, then by seasonal aspects; or alternatively, if the instrument type was most important, that could be the first criteria, etc. The classifications used are generally part of a pre-specified ontology, or common metadata vocabularies.

The primary objectives in this collaboration were to (1) provide ocean scientists with an improved data access interface when accessing datasets served by BCO-DMO and (2) to provide the RPI semantic web graduate research program with direct contact with oceanographic data to extend existing data science work. The collaboration has focused on the development of an ontology to describe the relationships that exist between different search classifications important to the BCO-DMO data holdings. This ontology will be incorporated into a smarter front-end for on the BCO-DMO website to enable faceted searching.

One potential benefit for WHOI oceanographers emerging out of the BCO-DMO collaboration is the possibility of a common data access framework that will make a large number of data-intensive oceanographic data accessible through the same interface, thereby increasing the interoperability of the data sets. This will also encourage additional interdisciplinary oceanographic and ecological science investigations.

In addition, RPI plans to merge the eventual ontology for this project with several other domain-specific ontologies it has already built, so that the resulting ontology will be applicable in many new science areas.

Relevant Staff:

BCO-DMO: Cyndy Chandler, Bob Groman, Peter Wiebe, and Dave Glover

Ocean Informatics Working Group: Peter Fox, Art Gaylord, Andrew Maffei, and Jennifer Schopf

RPI Staff: Peter Fox, Patrick West, Stephan Zednik

Last updated: April 7, 2010