Notes on Combined discussion of VoCamp, USGS Ontology for The National Map, Populating a Geospatial Ontology repository and Coordination on Integrated Demos (27GJ)
(As compiled by Gary Berg-Cross) (27GK)
Summary (27H9)
Following discussion of vocabulary development and publically available data from USGS and others, a 2 track plan was suggested for moving forward within the coming year. The main direction is to use the new, publically available data from USGS data in RDF using existing models as it becomes available in Dec. The data would be used in a Service hosted at USGS and available to the public. Two additionally things needed are an agreement on the Use Case (what to demonstrate in the service) and who would build the front end GUI. (27GM)
The other track is to pursue the possibility of using Google resources. The group looked to SOCoP to summarize the proposal and initiate activities towards development of such a service in cooperation with others. (27GO)
Details (27H6)
Todd Pehle (Orbis Technologies) kicked off the meeting by summarizing the recent VoCamp (http://vocamp.org) held after ISWC to develop a small, core, lightweight set of vocabularies based on neo geography. His slides are available discussing the possibility to build on the linked data thrust of Sir Tim Berners-Lee to put large amounts of public data on the Web." Todd cited the growth of location/spatial information on the web and noted that its use would be improved by a proper model to guide us in the use of such information. Vocabulary development is part of that model development effort. There were a number of people such as Jens Lehmann who were enthusiastic about having features as referencable URIs. (27GX)
Gary Berg-Cross (Knowledge Strategies) proposed that we build on this work with an integrated demo and publically available national map data. (27GY)
Luis Bermudez (SURA) thought that The National Map integration of data (between now and 2015) provided a new opportunity. We have lots of data, including for events and there are low-hanging fruit opportunities if we build relations, as noted by Dalia below. (27GZ)
Dalia Veranka (USGS) believes that we cant advance without the type of activity (e.g. Geographic Feature Type Ontology ) that Todd outlined in his talk. She briefed several slides on todays The National Map data layers base map and how it goes beyond the Ordinance Survey by including environmental and human use information. There is a new, integrated set of hydrology data stored in Oracle and represented in OWL. This set of data for 6 watersheds and 2 urban areas (St. Louis and Atlanta) will include all 8 data layers and become available for the watershed area in December. (27H0)
(see cegis.usgs.gov/ontology.html) (27H1)
These are more sophisticated and go beyond the base map by the use of Classes for GeoNames etc. There are no axioms yet and more relations are needed, but it is a first step as a contribution to linked data. The first conversion of data is being kept in house, but USGS will work with people after full conversion. That is, the converted data sets have been approved for people to work with. (27H2)
Carl asked how we can learn more about the various data layers. Dalia noted that information is on the site (see Dalias handout). This describes the data models for the features such as topographic features. (27GQ)
The internal USGS research group is small, but has a new server for its work. Logistics for this should be accomplished by the end of Dec. Troy Self (BBN) asked if the Oracle data was in RDF. Dalia said that isnt in RDF yet, but should be in a few weeks. (27GR)
Carl asked if any mashups of TNM data were available. Dalia noted that these were difficult up to now without an RDF form. There will be a public announcement Dec. 3 in Reston. Lynn promised to send a link for a viewer that is available now. Dave Kolas agreed that getting a core vocabulary of geo and relations is a start, but just as important to get useful data out along with tools to use that data. The data/knowledge representation that gets used for this may not be the best (but we can live with this for a while). (27H3)
Open Discussion of Publically Available Data (27H7)
Carl Diegert (Sandia) Noted that Mike Jones (now of Google who bought Keyhole, his satellite image company who owns data) proposes that we give people tools and people will then build high quality maps (of course Google will then own them). A related idea on tools was presented last year by Todd on a collaborative Wiki and the use of DBPedias location data (see http://wiki.dbpedia.org/Datasets). We would need the data to be in a good form to allow spatial queries. Another source that Todd noted is Data.gov which has been converted to RDF (Jim Hendler and Li Ding at RPI) see http://blogs.zdnet.com/semantic-web/?p=319 (27GS)
Discussion of an Integrated Demo (target is within the coming year 2010) (27H8)
With the availability of converted National Map data for select watershed and urban areas Luis thought this a good platform for testing within the Semantic Web paradigm of linked data. He proposed using at least 2-3 features from TNM for the available watershed data as a way of going forward. Lynn noted that USGS has a goal of small demos within FY 2010!!! (27GT)
Luis noted that atmospheric data for integration is publically available to be linked to it (check University source perhaps a look at http://www.wrclib.noaa.gov/lib/env_data/atmospheric.html will tell us.) But this integration requires handling dynamic data and Carl wondered if our tools are up to handling this and making it publishable. Wed need a Web 3.0 approach, but this may be down the road and not needed now. (27GU)
The use of Encyclopedia of Life tagged data was suggested for an integrated demo. Moving forward on an integrated demo will require some modeling too. A key point is to make this MORE than a demo. The group, urged by Dave Kolas, agreed that we should create a living service. (27H4)
Carl asked who the customers for such a service would be and whether SOCoP could get a dedicated server, perhaps donated and keep this in operation. Luis asked if Google might provide Cloud resources to begin with. Gary asked if the new USGS server would be available for such use and Lynn said that it would be available in the future. (27GV)
The Plan (27GW)
A 2 track plan was mentioned going forward in both the USGS direction and pursuing the possibility of using Google. We had more discussion of the use of USGS data in RDF using existing models when it becomes available. Two things needed are an agreement on the Use Case (what to demonstrate) and who would build the front end GUI. Carl noted that if someone pitches in early they will have more influence on how the RDF is used. The group conclusion was a resounding interest in such an integrated demo given no restrictions in making test data available as a service for public use. (27GN)