Evolving the BCO-DMO search interface - experience with semantic and smart search Cyndy Chandler (WHOI) Peter Fox (RPI and WHOI) Robert Groman, Dicky Allison Andy Maffei (WHOI) Patrick West, Stephan Zednik (RPI) EGU 2010 Ocean Informatics

Basis of effort Staff and graduate students from the Tetherless World Constellation at Rensselaer Polytechnic Institute (RPI) have been collaborating with the Biological and Chemical Oceanography Data Management Office (BCO-DMO) -- a project operating out of the Woods Hole Oceanographic Institution and funded by the National Science Foundation. RPI staff and BCO-DMO team-members have been working with oceanographers, data managers, ontology modelers, software engineers and other experts to iteratively design and develop a semantically enabled prototype showing how domain scientists are able to perform better and smarter searches for data, access and manipulate more data sets, and begin to keep track of data provenance. There are plans for the features demonstrated in this prototype to be incorporated into BCO-DMO’s production website. If time: image informatics.. New results Tetherless World Constellation 2

3

4

5

6

7

Modern informatics enables a new scale-free** framework approach Use cases Stakeholders Distributed authority Access control Ontologies Maintaining Identity

Team… Collaboration: Small team of mixed skills created in order to provide a scientific infrastructure that is usable and extensible, providing semantic integration, and knowledge representation while requiring depth in each of the science areas. Facilitator - knows iterative methodology, guides the exercise Domain experts – knows resources, data, applications, tools Ontology modelers – to extract objects/relations from use cases and discussion Data Managers – understands the storage, organization and access to datasets Software engineers – responsible for architecture and technology aspects Scribe – capturing everything discussed Social Scientist – optional, as process is as much a social exercise as it is a technical and methodical activity Tetherless World Constellation 9

Tools Omni Graffle – Creation of Faceted-Browse Mockups CmapTools COE – Creation of Ontology Models, Causality graphs for provenance Protégé – Creation of Ongology and Individuals Skype (IM and VOIP), Dimdim (Web Conferencing), MediaWiki – Collaboration tools Google Web Toolkit + SmartGWT – Rapid UI Prototyping Jena/TDB and Joseki – triple store and SPARQL endpoint server – can be extended to perform reasoning and the execution of semantic rules. Tetherless World Constellation 10

Use cases 1.Do you have any data online from Hutchins from award number OCE ? 2.I want to download (temperature, biological,...) data in the following areas (N. Atlantic, bounding box, where JGOFs survey was done,...) 3.What new data has been added since last year (and organize it by project) 4.Show me all the places where the surface temperature in the North Atlantic is 25 degrees during June. Tetherless World Constellation 11

Quick prototype of use case 1 Tetherless World Constellation 12

Evolving the ontology model Tetherless World Constellation 13

To… Example where the iterative process helped to develop an understanding by WHOI domain experts ontologies and translating their concepts into an ontology and the ontology developers to understand the specific domain vocabulary. Successive iterations helped to expand and simplify concepts and incorporate already existing ontologies. Similar in instrument, platform, parameter ontology development. Tetherless World Constellation 14 Includes all of the foaf concepts for name, contact information, interests

Current version Tetherless World Constellation 15

Current version Tetherless World Constellation 16

Summary Migrated a database driven, highly programmed implementation into an ontology and smart query driven search with modest effort (okay, a few brain cells died along the way) – Use case driven – Ontology driven at many levels – Application oriented, rapid prototyping All along the way, we evaluated our semantic developments (ontologies) and implementation to gauge their benefits or deficiencies Continuing to add functions based on new use cases Tetherless World Constellation 17

HABCAM Image Informatics Color and Illumination Prof. Chuck Steward (RPI) Students: Ryan Leary and Zack Schilling Problems addressed: – Illumination Across images Within image – Color Differing attenuation in water for red, green and blue – Demosaicing is noisy Approach: – Combined physical and empirical model

Color Correction Based on Beer’s Law Before After

Illumination Correction Based on Light-Field Map Before After Difference

Further Information Contacts: – – Tetherless World Constellation 21