Progress 07/15/01 to 09/30/11
Outputs
OUTPUTS: The most significant output from 2010 was the funding of the SOCoP INTEROP proposal submitted to the National Science Foundation in July 2009. The award was made Sept. 1, 2010. This grant funds six people in addition to N. Wiegand. Wiegand is the main Principal Investigator and coordinator. The work on this started in 2007 and finally reached a high level of fund-ability. Work was then needed to reduce the budget in July 2010. The next most significant output was the submission and support of an NSF REU (Research Experience for Undergraduates). This enabled several students to be hired and trained in research. Outputs also consisted of talks and papers and included conference calls for community building. The talks, papers, and other communication were examples of disseminating information about the project and the topics of work. As in past years, a talk was given at the Wisconsin Land Information Association (WLIA) conference to introduce local, state, and academic GIS professionals to new methods for achieving semantic interoperability in geospatial data. At the national level, the Spatial Ontology Community of Practice (SOCoP) continued its monthly conference calls involving N. Wiegand and had its annual workshop on Dec. 3, 2010. A talk was given remotely for that workshop. Other places in which a presence was made in 2010 regarding semantic interoperability were the CyberGIS workshop in Washington, D.C. in Feb. 2010 and the GEON cyberinfrastructure summer institute in San Diego in August 2010. Papers and abstracts were also written in 2010 for talks in 2011. An abstract submission for a talk was made for WLIA 2011. A talk abstract and session submissions were made to the American Association of Geographers (AAG 2011). A talk abstract was also made to ASPRS 2011. An accepted Cartographica paper was revised. No cost extensions were submitted twice. Students were hired for the REU. Work and travel were done for an NSF March proposal review panel. Finally, work was done to interact with a Digital Curation working group on campus to tie with the overall cyberinfrastructure goal of the NSF grants. PARTICIPANTS: Individuals: PI Nancy Wiegand Co-PI; Isabel Cruz, University of Illinois at Chicago, who works on complementary parts of the project; Partner organization: Computer Science Department, University of Illinois at Chicago; Collaborators: Director and staff at LICGF (UW-Madison), SOCoP (Spatial Ontology Community of Practice), Wisconsin Sea Grant, Training: About six undergraduates were hired to help with the project and learn about research TARGET AUDIENCES: This project is intended to serve geospatial professionals in the state and the nation. It will also help anyone, including the general public, who is looking for geospatial data. A target audience for hiring for the REU is undergraduates to enable them to become involved with research. This is part of a national initiative to engage students early in their careers to learn what research is and how it is done. Further, effort was made to hire women and minority undergraduates. So far, two women and one minority were hired along with three white males. Other target audiences for the work were GIS professionals in Wisconsin, other domain scientists and researchers at UW-Madison, and researchers at national level workshops and conferences. PROJECT MODIFICATIONS: Not relevant to this project.
Impacts
Again, the most significant outcome from 2010 was receiving an NSF INTEROP award for over 3 years. The award was the result of many outputs. Also, again, another outcome was the funding of an application submitted in April 2010 for an NSF REU grant. This allowed the hiring of a number of undergraduates to learn the research and contribute to the project goals. As stated last year, the overall goals of this project are to improve search, query, and integration of heterogeneous geospatial data. Work includes developing use cases, and, this year, a use case was explored for the Wisconsin Coastal Atlas project of Sea Grant, along with the beginnings of other use cases. Other outcomes include continued work by a student to create a land use ontology for Wisconsin and another student starting to put some sample Wisconsin data in RDF to serve it over the Web. This work has explored using various kinds of semantic technologies, such as Protege, RDF stores, SPARQL, Twinkle, and the AgreementMaker. N. Wiegand has also submitted a paper on a high level ontology design to be added to database management systems to handle the emerging complexity of types of data to be managed and related.
Publications
- Wiegand, N. Geospatial Cyberinfrastructure, DBMS, and Geospatial Ontologies. Position paper for National Science Foundation TeraGrid Workshop on Cyber-GIS, Feb. 2-3, 2010, Washington, D.C., http://www.cigi.illinois.edu/cybergis/participants.php
- Wiegand, N. 2010. Scenarios for Using Ontologies in Searching for Geospatial Data in Government Portals. accepted in DG 2010.
- Wiegand, N. 2010. Ontologies and Database Management System Technology for The National Map. Cartographica: The International Journal for Geographic Information and Geovisualization, 45(2):121-126.
- Wiegand, N., Kolas, D., and Berg-Cross, G. 2010. Intersecting Semantic Web and Geospatial Technologies. Guest Editorial, Transactions in GIS, April 2010, 14(2):93-95.
- Wiegand, N. Interoperability: Call for Engagement and Use Cases. WLIA 2011 talk abstract.
- Wiegand, N. INTEROP Network to Support Geospatial Data Semantic Interoperability. AAG 2011 talk abstract.
- Wiegand, N. Geospatial Semantics and Ontology. AAG 2011 session abstract.
- Wiegand, N. INTEROP Network to Support Geospatial Data Semantic Interoperability. ASPRS 2011 talk abstract.
- Wiegand, N. 2011. Using Semantic Technology to Organize and Relate Data Components. Submitted to a conference.
|
Progress 01/01/09 to 12/31/09
Outputs
OUTPUTS: Dissemination overview: SOCoP, Ontolog, WLIA talk, The National Map workshop, Terra Cognita workshop, and the Math/Bio Symposium. This last year continued to be a year of community building as well as moving forward on using ontologies in the geospatial area. Ontologies are useful for establishing a background knowledge base, enabling information organization, improving search, resolving heterogeneity, allowing for information integration, and generating new knowledge through inferencing. Outreach activities included exposing the geospatial community to the potential advantages of using ontologies and also to strengthen communication and coordination within and among researchers and practitioners interested in this new and evolving area. Reaching the State of Wisconsin geospatial community included another talk at the Wisconsin Land Information Association (WLIA) conference. And, coordinating with researchers was done in several ways. One is the growing connections of SOCoP members (Spatial Ontology Community of Practice) through continued monthly conference calls, writing a proposal together submitted July 23, 2009, and meeting in person at various conferences. Ontolog teleconference sessions also draw people together. In addition, N. Wiegand again proposed and was the main co-Chair of the interdisciplinary workshop, Terra Cognita, held in conjunction with the 2009 International Semantic Web Conference. This workshop brought Semantic Web researchers together with geospatial researchers. Also, N. Wiegand's invitation for attendance at the USGS Ontology for the National Map meeting in Feb. 2009, served to push various ideas forward regarding ontologies and geospatial data, as well as to further engage interested researchers. In addition, she presented a poster at the UW-Madison Math/Bio symposium in Nov. 2009. Papers and abstracts were also written. PARTICIPANTS: Individuals: PI Nancy Wiegand; Co-PI Isabel Cruz, University of Illinois at Chicago, who works on complementary parts of the project; Partner Organization: Computer Science Department, University of Illinois at Chicago; Collaborators: Director and staff at LICGF, contacts at WLIA (Wisconsin Land Information Association), SOCoP (the Spatial Ontology Community of Practice); Training: An undergraduate student was mentored A graduate student helped with the project. TARGET AUDIENCES: This project is intended to serve geospatial professionals in the state and the nation. It may also help anyone, including the general public, who is looking for geospatial data. Efforts included personal training and teaching in lab situations for the undergraduate mentee. Other efforts to explain, make aware, and disseminate the work were done through talks and posters at UW-Madison, at the State WLIA conference, and at a number of professional conferences and workshops. Further, to disseminate knowledge in this geospatial semantics area, N. Wiegand proposed a workshop at an international conference and organized and co-Chaired it. She also arranged for a special issue in a journal for the best papers from the workshop, carefully re-reviewed all papers, and wrote the guest editorial. PROJECT MODIFICATIONS: Not relevant to this project.
Impacts
The overall goals of this project are to improve search and resolve heterogeneities for geospatial data to enable data to be integrated and queried for user purposes. Research includes developing use cases, and, this past year, additional scenarios and sketches were developed as to how ontologies can help in the geospatial area, specifically for The National Map. The impacts this year included a strengthening both of people being made aware of the potential of ontologies and people coming together to help coordinate efforts in research and practice. An example outcome showing the effect of getting the geospatial community to understand and appreciate ontologies is that someone at UW-Madison included N. Wiegand to incorporate ontologies in a geospatial Sea Grant proposal. Impacts were also achieved through increased numbers of conference calls, travel to meetings and conferences, and organizing a second geospatial/semantic web Terra Cognita workshop at an international conference. Also, the SOCoP group continues to get stronger, and they put on another fall workshop. With SOCoP, N. Wiegand submitted an NSF proposal. The outcome of the Terra Cognita workshop was refereed papers published in the journal Transactions in GIS with a guest editorial by N. Wiegand. An additional outcome is the continuation of the investigation of how geospatial ontologies might be used in new ontology-enhanced database management systems. Suggestions for DBMS modeling and system improvement were also pushed forward in 2009 although more needs to be done before a paper can be submitted on this. Other work included developing an algorithm for using ontologies in searching for geospatial data. N. Wiegand continues to gain and push forward knowledge in this area. She carefully reviewed all papers submitted to the Terra Cognita workshop, which she organized, and also re-reviewed the revised editions for the TGIS journal.
Publications
- Wiegand, N. and Kolas, D. 2008. Geospatial Enhancements for the Semantic Web, Guest Editorial, Transactions in GIS, Vol. 12, Issue s1, pp. 1-3.
- Wiegand, N. 2008. A Geospatial Semantic Web Scenario, Technical Report.
- Wiegand, N. 2009. Improving Search for Geospatial Data Using Ontologies, Technical Report.
- Wiegand, N., Kolas, D., and Berg-Cross, G. 2009. Intersecting Semantic Web and Geospatial Technologies, editorial, Special issue of Transactions in GIS for Terra Cognita 2009.
- Wiegand, N. 2009. Ontologies and DBMS Technology for The National Map, Submitted.
- Wiegand, N. 2009. Meta-Schema Ontology, in progress.
|
Progress 01/01/08 to 12/31/08
Outputs
OUTPUTS: This past year was a year of community building as well as moving forward on using ontologies in the geospatial area. Ontologies are useful for establishing a background knowledge base, enabling information organization, improving search, resolving heterogeneity, allowing for information integration, and generating new knowledge through inferencing. Outreach activities included exposing the geospatial community to the potential advantages of using ontologies and also to strengthen communication and coordination within and among researchers and practitioners interested in this new and evolving area. Reaching the geospatial community was mainly done through state conference talks. And, coordinating with researchers was done in several ways. One is the growing connections of SOCoP members (Spatial Ontology Community of Practice) through continued monthly conference calls, writing a proposal together, and meeting in person at various conferences. N. Wiegand proposed and was the main co-Chair of an interdisciplinary workshop, Terra Cognita, held in conjunction with the 2008 International Semantic Web Conference. This workshop brought Semantic Web researchers together with geospatial researchers. N. Wiegand's invitation for attendance at the USGS Ontology for the National Map meeting also served to push various ideas forward regarding ontologies and geospatial data, as well as to further engage interested researchers. Other work included developing a geospatial scenario to illustrate the potential use of ontologies. This was presented in a talk at the Semantic Technologies Conference and also formed part of a paper. Additional work was done on the use and modeling of ontologies in Relational Database Management Systems, specifically Oracle. A sketch was presented as to how ontologies in Oracle could be useful to the geospatial community and to identify some needs that are not yet met in Oracle. Further work will be done to enhance current RDBMS design regarding ontologies. Also, in Spring 2008, an undergraduate student was mentored as part of Undergraduate Research Scholars program. In addition, numerous posters were presented as well as a number of talks. Papers and abstracts were also written. PARTICIPANTS: Individuals: PI Nancy Wiegand; Co-PI Isabel Cruz, University of Illinois at Chicago, who works on complementary parts of the project; Partner Organization: Computer Science Department, University of Illinois at Chicago; Collaborators: Wisconsin State Cartographer's Office, Director and staff at Land Information and Computer Graphics Facility (LICGF), contacts at WLIA (Wisconsin Land Information Association), SOCoP (the Spatial Ontology Community of Practice); Training: An undergraduate student was mentored and a graduate student helped with the project. TARGET AUDIENCES: This project is intended to serve geospatial professionals in the state and the nation. It may also help anyone, including the general public, who is looking for geospatial data. Efforts included personal training and teaching in lab situations for the undergraduate mentee. Other efforts to explain, make aware, and disseminate the work were done through talks and posters at UW-Madison, at a regional Database conference, at the State WLIA conference, and a number of professional conferences and workshops. N. Wiegand organized and co-Chaired a workshop at an international conference. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.
Impacts
The overall goals of this project are to improve search and resolve heterogeneities for geospatial data to enable data to be integrated and queried for user purposes. This past year, additional scenarios and sketches were developed as to how ontologies can help in the geospatial area. The impacts this year included a strengthening both of people being made aware of the potential of ontologies and people coming together to help coordinate efforts in research and practice. Both types of impacts were done through increased numbers of conference calls, travel to meetings and conferences, and organizing a geospatial/semantic web workshop. The outcome of the workshop was referred papers published in the journal Transactions in GIS with a guest editorial by N. Wiegand. An additional outcome is the investigation of how geospatial ontologies might be used in new ontology-enhanced database management systems. Suggestions for modeling and system improvement were also made although more needs to be done. As to a change/increase in knowledge over the year, N. Wiegand performed thorough background work to be able to present at the 2008 Semantic Technologies Conference. Much was also learned to prepare for two talks for the 2008 GIScience Conference, including a talk on service-oriented architecture. Also, there was a large knowledge gain in reviewing all papers submitted to the Terra Cognita workshop, which she organized.
Publications
- Wiegand, N. 2008. Improving Search for Geospatial Data Using Semantic Web Technologies. (pending).
- Wiegand, N. 2008. A Geospatial Semantic Web Scenario. (pending).
- Wiegand, N. 2008. Exploring RDBMS Support for Ontology Enhanced Searching in Geospatial Portals, GIScience 2008, Park City, Utah, pp. 196-200. (published).
- Wiegand, N. 2008. Ontology Enhanced Databases. Position paper accepted to the UCGIS Specialist Meeting on Ontology for the National Map, Feb. 3-4 Washington, D.C. (in press).
- Wiegand, N. and Kolas, D. 2009. Geospatial Enhancements for the Semantic Web, Guest Editorial, Transactions in GIS. (in press).
|
Progress 01/01/07 to 12/31/07
Outputs
OUTPUTS: Our work continued the effort to improve searching for geospatial data. Here, we consider a cyberinfrastructure consisting of geospatial portals on the Web. As part of a task-based approach, we propose using a place ontology. A place ontology will be useful in general for many purposes. A place ontology contains more information and has a different structure compared to a gazetteer. A gazetteer is usually a table with each entry describing a place. Information includes the type of place and the city, county, and state in which the place is located, along with the x,y coordinates of the place. A place ontology is modeled as a graph with object properties between places and jurisdictions that explicitly show composition (part_of) and located_in relationships. Compared to a gazetteer, additional spatial type relationships can also be modeled such as near, adjacent, north of, and so on. In addition, further associations/relationships can be modeled with other types of data as
they are related to places. Our idea of using a place ontology to improve search is that each geospatial data source will be directly linked to a place in a place ontology. This removes the inaccuracy that results from using x,y bounding box coordinates to determine the range of a data source and also removes the ambiguity from literally comparing a metadata place keyword to a gazetteer entry (i.e., many places have the same name). A place ontology also allows query expansion using transitive relationships. We began modeling a place ontology thinking about Wisconsin as an example. We are using Protege as an OWL graphical editor for the model. However, Protege has limited query support and currently does not support transitive relationships, which are critical to reasoning over a place ontology. In this regard, we are exploring SPARQL, an RDF/OWL query language. Over this year, we experimented with various models of places and the description logic paradigm of Protege. Two
undergraduate students were mentored on this project. Several talks have been given on this work and posters were presented to state, university, and national audiences. Papers have been submitted.
PARTICIPANTS: Individuals: PI Nancy Wiegand Co-PI Isabel Cruz, University of Illinois at Chicago, who works on complementary parts of the project. Partner Organization: Computer Science Department, University of Illinois at Chicago. Collaborators: Wisconsin State Cartographer's Office, Director and staff at LICGF, contacts at WLIA (Wisconsin Land Information Association). Training: Two undergraduate students were mentored. A graduate student helped with the project.
TARGET AUDIENCES: Target: This project is intended to serve geospatial professionals in the state and the nation. It may also help anyone, including the general public, who is looking for geospatial data. Efforts included personal training and teaching in lab situations for the two undergraduate mentees. Other efforts to explain, make aware, and disseminate the work were done through talks and posters.
Impacts
The impact of this work is to use Semantic Web technologies to help locate and query geospatial data. Geospatial data are distributed over the Web and/or accessible through portals but are currently difficult to locate due to heterogeneous metadata terms as well as the large variety of types of data. One goal is to enable more precise searching of geospatial data sources using ontologies, for example. Using ontological concepts instead of key words enables a user to locate data sources even if the descriptive metadata has heterogeneous terms. We are also trying a task-based approach to help formalize and specify which types of data are needed for particular tasks. A place ontology could be part of a task-based approach. A place ontology would be useful in many applications. Contrary to a gazetteer, nodes in a place ontology could be linked to many other types of data. With ontologies, rules, and an inference engine, a Web-based geospatial information system could
eventually automatically locate and query data sources. This would be extremely helpful in emergencies. Through this work, we have made geospatial professionals aware of the potential benefit Semantic Web technologies may have on searching for and querying geospatial data. One graduate student and two undergraduate students have been introduced to the research process in general and this project in particular. A further proposal has been written to continue and expand the work. Many people contacts were made university-wide, state-wide, and nationally to pull ideas and expertise together.
Publications
- Wiegand, N. and Garcia, C. 2007. A Task-Based Ontology Approach to Automate Geospatial Data Retrieval, Transactions in GIS Special Issue on the Geospatial Semantic Web, 11(3):355-376.
- Wiegand, N. 2007. Semantic Web for Geospatial E-Government Portals, 8th Annual International Conference on Digital Government Research, May 20-23, 2007, Philadelphia, PA, pp. 298-299.
|
Progress 01/01/06 to 12/31/06
Outputs
Our project began by focusing on methods to improve searching for geospatial data. Although geospatial portals have potential for finding geospatial data, currently, a user's keyword search may not return the appropriate data or may return numerous results, many of which are not relevant. One of our approaches to improve searching for data is to add task-based knowledge to geospatial portals. To do this, first, a taxonomy of types of tasks is needed, and we began by using the Homeland Security Working Group's symbology list for emergency events. We also investigated the kinds of data sources that are needed in emergencies and for other types of tasks. We took several approaches to obtain this information. We entered into a dialog with the Wisconsin Emergency Management Task Force, a local committee actively engaged in the development of emergency data protocols. We also contacted people via email (e.g., the Forest Service) and used a Land Use Planning Manual available
for Wisconsin. Although we collected sample information, we found that we needed additional ontological information that is not now readily available. That is, for a task-based approach, in addition to needing a full ontology of tasks, we need an ontology of data sources. Furthermore, we need knowledge as to what data sources are needed for which tasks. This ontological information can be very valuable in general. For example, a full ontology describing types of geospatial data sources would codify knowledge that now only resides with a limited number of geospatial specialists. Also, codifying types of data needed for types of tasks will facilitate nonspecialists' ability to find data. We are particularly interested in further pursuing development of ontologies for Homeland Security and Emergency Preparedness. Simultaneously with investigating geospatial data and tasks, we began experimenting with Semantic Web technologies, such as Protege OWL. The purpose was to see if these new
technologies could be used to improve searching for geospatial data and, conversely, also to see if we could recommend improvements to the technology based on geospatial needs. We experimented with Semantic Web technologies and languages such as Jess (a rule engine), JessTab (a plugin that converts OWL code to Jess assertions), and SWRL (a W3C proposed Semantic Web Rule Language). A large part of our time this year was spent in developing a prototype task-based software system using Protege OWL, Jess, and JessTab. Jess was needed to augment Protege with a rule engine. The software development took more time than we anticipated because of the complexity of modeling, limitations of the Protege Editor, and learning Jess. Also, because of modeling and JessTab limitations, we performed a series of pilot studies and Protege project versions to evaluate different models and methods to get our prototype to work. We had success in the end. Our final system takes locations and type of task as
user input and automatically returns the URLs of needed data sources. To be fully deployable, however, more in-depth ontological information is needed.
Impacts
The impact of this proposal is to bring closer to reality semantic integration over diverse geospatial data. We intend to provide methods and tools to support semantic integration that will enable analysis and decision-making for various applications such as emergency response or land use planning. One goal is to enable more precise locating of geospatial data sources distributed over the Web. The use of metadata integration methods and ontologies will help accomplish this goal. Another goal is to resolve diverse geospatial data models to be able to integrate heterogeneous data. And, our work strives for the ability to query data over the Web, which again involves resolving semantic heterogeneity. Our vision is for a Web-based geospatial information system that uses ontologies. This will enable a user to precisely locate data sources even if they were described in heterogeneous metadata terms. The user would also be able to query over the content of multiple
heterogeneous data sources using integrated data models or schemas. Furthermore, using the ontology along with rules, a Web application may be designed that would automatically locate and query data sources.
Publications
- Wiegand, N. 2006. Using Semantic Web Modeling and Technology for Automated Search for Data.. Abstract for Geoinformatics Conference, May 10-12, 2006, USGS, Reston, VA, http://www.geongrid.org/geoinformatics2006.
- Wiegand, N. 2006. A Semantic Web to Automate Searching for Geospatial Data. dg.o2006, 7th Annual International Conference on Digital Government Research, May 21-24, 2006, San Diego, CA, pp. 464-465.
|
Progress 01/01/05 to 12/31/05
Outputs
The objective of this proposal is to provide semantic integration over diverse geospatial data to enable locating and querying data sets distributed over the Web. Work includes metadata integration methods to enable more precise location of data sources as well as resolving diverse geospatial data models. Our goal is to provide methods and tools to support the semantic integration of geospatial data to enable analyses and decision-making for various applications such as emergency response. Our vision is for a Web-based geospatial information system that uses ontologies so that 1) a user can precisely locate data sources that were described in heterogeneous metadata terms, 2) the user can query over the content of multiple heterogeneous data sources, and 3) ultimately, a Web application can be designed to automatically locate and query data sources using the ontology methodology. Results so far (6 months): We are developing an ontology model for a prototype system that
attempts to enable automatic location of geospatial data sources for particular tasks. We assume that types of data sources needed for a specified task can be delineated by domain experts and that the types of data sources needed for a task remain the same across various locations. If so, reasoning can be done to find data source instances for a task at a particular location. The model is being prototyped using the Protege OWL Ontology System from Stanford. We are supplementing the system with a plug-in for a rule-based inference engine (Jess from Sandia National Laboratories). The inference engine is needed for inferencing over the OWL ontology using rules defining constraints needed to determine data sources.
Impacts
The impact of this proposal is to bring closer to reality semantic integration over diverse geospatial data. We intend to provide methods and tools to support semantic integration that will enable analyses and decision-making for various applications such as emergency response or land use planning. One goal is to enable more precise locating of geospatial data sources distributed over the Web. The use of metadata integration methods and ontologies will help accomplish this goal. Another goal is to resolve diverse geospatial data models to be able to integrate heterogeneous data. And, our work strives for the ability to query data over the Web, which again involves resolving semantic heterogeneity. Our vision is for a Web-based geospatial information system that uses ontologies. This will enable a user to precisely locate data sources even if they were described in heterogeneous metadata terms. The user would also be able to query over the content of multiple
heterogeneous data sources using integrated data models or schemas. Furthermore, using the ontology along with rules, a Web application may be designed that would automatically locate and query data sources.
Publications
- Wiegand, N. 2005. A Geospatial Semantic Web for Searching, In roceedings First International Conference on GeoSpatial Semantics, Nov. 29-30, 2005, Mexico City, Mexico.
- Wiegand, N. 2005. Searching for Geospatial Government-Produced Data, In Proceedings National Conference on Digital Government Research, dg.o2005, May, pp. 269-270.
- Wiegand, N.; Cruz, I.F.; Zhou, N.; and Sunna, W. 2005. A Web Query System for Heterogeneous Government Data, International Journal of Electronic Government Research, Vol. 1, Issue 2, pp. 63-81.
- Wiegand, N. and Zhou, N. 2005. Ontology-Based Geospatial Web Query System, P. Agouris & Croitoru (Eds.), Next Generation Geospatial Information: From Digital Image Analysis to Spatio-Temporal Databases, ISPRS Book series,, Balkema, Taylor & Francis.
|
Progress 01/01/04 to 12/31/04
Outputs
The final outcome of this research is an enhanced Internet XML Database Management System (DBMS) that supports querying and semantic integration over diverse and distributed government data. The motivation for the project was a proposed Wisconsin Land Information System, which would allow distributed government-produced geospatial data sources to be available from a Web portal. We extended a search and download vision by adding a potential query component over data source content. To be able to query heterogeneous data, however, we introduced ontologies and ontology mapping. The significance of this project is first that it illustrates the potential for full query power being available over the Web so that a user can retrieve answers to questions, rather than just obtaining URLs of data sources, as occurs with current HTML Web search engines and existing geospatial portals. Furthermore, because distributed Web data produced by different government jurisdictions are
heterogeneous, we developed a method to mediate between diverse values. We use ontology mappings and enhancements to the XML-QL query language. Our work extends research in the areas of Database Management, XML technologies, the Semantic Web, and geospatial data. In particular, we extend existing DBMS research to resolve heterogeneity at the value level in addition to the schema level. That is, for a particular attribute (field), we map between values used in different data sets. We illustrated our system using Land Use data because of its heterogeneity across multiple jurisdictions. In our software system, a user wanting to find a certain land use that covers an area over several counties can just specify the counties and the land use. Our software will automatically map between the stated land use and the various terms used by the individual counties. This project has resulted in significant outreach through talks and demos given at local, state, and national conferences. Audiences
included local and state officials and researchers in Digital Government, Computer Science, Remote Sensing, the Semantic Web, and Geographic Information Science. The work has helped promote interdisciplinary research and dialog. There are two main software products from the current work. The first is an operational Internet XML Database Management System enhanced with an ontology subsystem that is able to mediate over heterogeneous government-produced data to produce query results. The system also delivers enhanced results that are summarized, annotated, and spatially displayed. We started from an existing prototype XML Internet DBMS (Niagara) to which we added an ontology subsystem and other enhancements. The other product is a stand-alone tool developed at the University of Illinois at Chicago to semi-automatically create mappings between a global ontology and a local schema. This tool displays global ontology terms as well as local terms to enable a domain expert to map between the
two. Some automatic mapping of terms is done initially to simplify the task. The tool automatically generates an agreement file containing the mappings, which are then used in the Web query system to generate subqueries.
Impacts
The impact of this project is improved searching for data over the Web and the ability to query and synthesize distributed Web data, which will result in enhanced decision-making. This project impacts various types of Web portals including the vision for a Wisconsin Land Information System. We demonstrated the potential to extend capabilities of portals that currently use keyword searches and return URLs of data sources. Enhanced capability includes full Database Management System query power in which answers to questions can be returned, assuming data are marked up in XML. Such query power can also be used to enhance the search process to more precisely locate data sources. Furthermore, we demonstrated a method to resolve heterogeneous terms used by various government jurisdictions. That is, a user will be able to issue a query over the Web that covers multiple jurisdictions or even the entire state. The user will receive an answer even though multiple, distributed,
and heterogeneous data sources are involved. We developed an operational framework and architecture for such a system.
Publications
- Wiegand, N. Querying Metadata Over the Web to Locate Geospatial Data, In Proceedings of GIScience 2004 Extended Abstracts and Poster Summaries, October 2004, pp. 224-226.
- Wiegand, N.; Zhou, N.; Cruz, I.F.; and Sunna, W. 2004. A Web Query System for Heterogeneous Government Data, In Proceedings of the National Digital Government Conference, May, Full paper, pp. 27-36.
- Wiegand, N.; Zhou, N.; Cruz, I.F.; and Sunna, W. 2004. Ontology-Based Geospatial XML Query System, In Proceedings of the National Digital Government Conference, May, 2004, Demo and paper, pp. 289-290.
- Wiegand, N. 2004. Integrating Metadata Development, XML, and DBMS Search and Query Techniques in a State of Wisconsin Land Information System, In Proceedings of the National Digital Government Conference, May, 2004, Project Highlights paper, pp. 367-368.
- Wiegand, N. 2004. Internet Querying of Metadata to Search for Geospatial Government Data, In Proceedings of the National Digital Government Conference, May, 2004, Poster and paper, pp. 369-370.
- Wiegand, N. 2004. Being Successful in Digital Government Projects, In Proceedings of the National Digital Government Conference
|
Progress 01/01/03 to 12/31/03
Outputs
We continued research to provide full Database Management System (DBMS) query support over data that will be part of a Wisconsin Land Information System. As a review of our project, the main focus of our work is data heterogeneity. We started our work focused on the semantic problem between land use coding systems, which are important for comprehensive planning. We analyzed the types of semantic discrepancies between various land use code systems and formulated a semi-automatic method to relate between codes. We use an ontology-based approach in which local experts map their code systems to a master set of codes (the ontology). Diverse code systems can then be compared by their expression in the ontology terms. We have a working XML Internet DBMS to which we added an ontology subsystem. Our user interface obtains geographical information along with a land use code predicate. Our system automatically resolves the differences in land use codes over the selected
geographical areas by performing lookups in stored agreement files. In addition to returning the data records satisfying the land use code selection over the combined geographic areas, we output the semantics of the code mappings to the user. As stated, our Internet DBMS uses an XML agreement file to look up comparisons between the ontology and each land use code system. These files are produced by a computer tool (designed by our subcontract) in which a local expert maps the ontology land use code to each local code. In the last year, we added more automation to the tool. Now, an automated reasoning system maps the ontology to local codes whenever possible. As a result, the local expert only has to map remaining unresolved codes. We also re-designed our user interface to match that of new geospatial portals emerging from the Federal government. In addition, we made it easier to express more complex queries in our Internet DBMS rather than merely selecting a land use code as a
predicate. We fine-tuned our demo. Our computer programming efforts are in Java. Much of the last year was spent publishing our accomplishments and giving talks and demonstrations at many conferences at the local, state, and national levels. A number of reports on our work were also done. Our results are significant because so much information on the Web cannot be used because of the differences in expression for similar concepts used by each jurisdiction and data publisher. We are working to resolve such differences automatically to support decision-making across jurisdictions, for example.
Impacts
Our vision is that users of a Wisconsin Land Information System will be able to issue a query over the Web that would cover multiple jurisdictions or even the entire state. The user would receive an answer even though multiple, distributed, and heterogeneous data sources are involved. To enable this, an Internet Database Management System (DBMS) is needed, and there are emerging prototypes. However, more than this, an Internet DBMS would not only have to automatically find the data sources containing answers to a query; it would have to automatically resolve differences in terms and meanings between the diverse data sets. We are working on semantic integration tools and adding them to a prototype Internet DBMS. We have developed a framework and architecture for this vision.
Publications
- [CR03a] Cruz, I. F. and Rajendran, A. "Semantic Data Integration in Hierarchical Domains", IEEE Intelligent Systems, Vol. 18, No. 2, March-April, 2003, pp. 66-73.
- [CR03b] Cruz, I.F. and Rajendran, A. 2003. "Exploring a New Approach to the Alignment of Ontologies", Workshop Semantic Web Technologies for Searching and Retrieving Scientific Data, in cooperation with the International Semantic Web Conference, Sanibel Island, Florida, pp. 7-12.
- [Wie03] Wiegand, N. 2003. Book review for (Spatial Databases with Application to GIS by Philippe Rigaux, Michel Scholl, and Agnes Voisard, Morgan Kaufmann, 2002), SIGMOD Record, Vol. 32, No. 4, December, pp. 111-112.
- [Wie03] Wiegand, N. 2003. "Searching for Geographic Data Sources for Land Use Planning", Semantic Web Technologies for Searching and Retrieving Scientific Data Workshop at the International Semantic Web Conference (ISWC), Sanibel Island, Florida, October 20, 2003, Poster, pp. 93-95.
- [WZ03] Wiegand N. and Zhou, N. 2003. "Ontology-Based Geospatial Web Query System", International Workshop on Next Generation Geospatial Information (NG2I), Boston, Oct. 19-21, 2003, Published.
- [WZC03] Wiegand, N.; Zhou, N.; and Cruz, I.F. 2003. "A Web Query System for Heterogeneous Geospatial Data", Scientific and Statistical Database Management, SSDBM, July, pp. 262-265.
- [WZC+03a] Wiegand, N.; Zhou, N.; Cruz, I.F., and Ventura, S. 2003. "Extending XML Web Querying to Heterogeneous Geospatial Information", In Proceedings National Conference on Digital Government Research, dg.o2003, May, pp. 279-282.
- [WZC+03b] Wiegand, N.; Zhou, N.; Cruz, I.; and Sunna, W. 2003. "Resolving Schema and Value Heterogeneities for XML Web Querying", Semantic Integration Workshop at the International Semantic Web Conference (ISWC), Sanibel Island, Florida, October 20, Demo, pp. 149-153.
- [WZV+03] Wiegand, N.; Zhou, N.; Ventura, S.; Cruz, I.F.; and Sunna, W. 2003. "Resolving Schema and Value Level Heterogeneities for XML Web Querying", Demo, In Proceedings National Conference on Digital Government Research, dg.o2003, May, p. 316.
- [Zho03] Zhou, N. 2003. "A Study on Automatic Ontology Mapping of Categorical Information", In Proceedings National Conference on Digital Government Research, dg.02003, May, Boston, pp. 401-404.
|
Progress 01/01/02 to 12/31/02
Outputs
We are continuing research to provide full Database Management System (DBMS) query support over data that will be part of a Wisconsin Land Information System. The main focus of our work is data heterogeneity. We are working on the semantic problem between land use coding systems, which are important for comprehensive planning. We analyzed the types of semantic discrepancies and formulated a semi-automatic method to compare codes. We use an ontology-based approach in which local experts map their code system to a master set of codes (the ontology) developed by domain experts. Diverse code systems can then be compared by their expression in the ontology terms. Our initial ontology only uses general categories of land use codes (e.g., residential, commercial, agriculture), but we improved our methods to be able to have as much detail as needed including any level of subcategory. Moreover, if necessary, we allow the ontology to be a super set of coding systems rather than
a master code scheme per se. Then, contrary types of categories can co-exist in the ontology. We developed a computer tool (done by our subcontract) to help each local expert map the ontology land use code to each local code set using mapping relationships 1:1, 1:N, N:1, and 1:NULL. Other semantic information is included using XML tags. The tool automatically stores the mappings by creating a computerized XML agreement file for each local code set. We developed a query rewrite system to generate subqueries to send to a base XML search and query engine (Niagara). Another innovation is a Domainspace statement that extends the XML-QL query language to formally express the unique type of query we are handling which has the same predicate ranging over multiple jurisdictions (data sets). Using information from the Domainspace statement and the agreement files, we automatically resolve the ontology land use code chosen by the user to appropriate local codes. We have a working demo that does
relational selection when the user specifies a land use code and jurisdictions. We wrote a new interface client (GUI) for Niagara and added our ontology and query rewrite subsystem. The computer programming effort is in Java. In addition to returning the data records satisfying land use code selection, we output the semantics of the code mappings to the user. To summarize, our innovations in the past year were to: design a method to do precise land use code mappings using an ontology approach, develop a computer tool to semi-automate the mappings and automatically store the mappings in agreement files, extend an XML query language with a Domainspace statement to handle our special type of multi-jurisdictional query, design a user interface, write a query rewrite system, and integrate all software into the Niagara Net DBMS.
Impacts
Our impact is an advance in methods of data integration. Data integration and interoperability are major subjects of research in diverse research communities (e.g., databases and artificial intelligence). Our method uses state-of-the-art technology (i.e., XML, ontologies, XML search and query engine). Also, our method enables data integration at the value level whereas prior work of others focused on the schema level. Developing a method for resolving semantic heterogeneity to be able to do DBMS type querying is at the forefront of research in this area. We are working on top of an existing XML query engine prototype so that our work can be readily transferred to commercial systems as they become available. From a practical perspective, we are addressing real world problems. Our work helps automate mappings between diverse land use coding systems. Our demo illustrates automatically finding records containing common land use codes over a multi-jurisdictional area. We
have disseminated our work by publications and national level talks.
Publications
- Wiegand, N.; Zhou, N.; Cruz, I.; and Rajendran, A. 2002. "Querying Heterogeneous Land Use Data Over the Web", GIScience 2002 Abstracts, M. Egenhofer and D. Mark (Eds.), Sept., Boulder, CO, pp. 207-210.
- Wiegand, N. "Investigating XQuery for Querying Across Database Object Types", SIGMOD Record, Volume 31, Number 2, June 2002, pp. 28-33.
- Wiegand, N.; Patterson, E.D.; Zhou, N.; Ventura, S.; and Cruz, I. 2002. "Querying Heterogeneous Land Use Data: Problems and Potential", In Proceedings National Conference on Digital Government Research, dg.o2002, May, pp. 115-121.
- Wiegand, N.; Zhou, N.; Patterson, E.D.; and Ventura, S. 2002. "A Domain Space Concept for Semantic Integration in a Web Land Information System", Demo paper, In Proceedings National Conference on Digital Government Research, dg.o2002, May, pp. 443-446.
- Cruz, I. F.; Rajendran, A.; Sunna, W.; and Wiegand, N. "Handling Semantic Heterogeneities Using Declarative Agreements", ACM GIS, Nov. 2002, pp. 168-174.
|
Progress 01/01/01 to 12/31/01
Outputs
Informal notification of our grant came in December 2000. We then worked to re-submit a revised budget with reduced funding by March 2001. On June 22, 2001, we held a successful government sponsor meeting at the Pyle Center to start our project. When official notification of our funding came at the end of July, position descriptions were posted to hire two graduate students. The first one started at the end of August, and the second started in September. The subcontract was re-done to accommodate Isabel Cruz's move to the University of Illinois at Chicago. Land use was chosen as the beginning theme for our project because of its importance and complexity. One student has become thoroughly knowledgeable about the types of land use codes used in Wisconsin. The other student has been working through the Java code of the Niagara XML query engine and making extensions to the system. We worked hard to write an extended abstract by the end of October for a potential demo. We
also developed an idea and submitted a short abstract for a talk at the 2002 WLIA conference. In addition to the government meeting, talks on the project have been presented to LICGF (Sept.) and to the Environmental Monitoring Seminar (Dec.). A great deal of reading and background work was done by all to be informed of related research and to work forward on developing new ideas. Among project members at UW-Wisconsin, group meetings are and have been held each week with many more individual meetings every few days. A meeting in Janesville was held Feb. 1 with Isabel Cruz and her graduate students from UIC. An overview of potential research areas was presented with a focus on the land use problems, and each group presented accomplishments so far. We are now working on papers, a demo, and talks describing our project for the Digital Government conference in May 2002.
Impacts
Our goal is to provide DBMS query support over Web documents that are part of the Wisconsin Land Information System (WLIS). Without our research contributions, WLIS would not incorporate full query support which has the potential of generating answers to questions rather than just URLs. We are working on the semantic resolutions necessary to query over heterogeneous data sets. Our initial focus is on land use codes. The impact of resolving land use codes across jurisdictions is the ability to do comprehensive land use planning over multi-jurisdictional areas.
Publications
- Wiegand, N. 2002. "Investigating XQuery for Querying Across Database Object Types", SIGMOD Record, To appear.
- Wiegand, N.; Zhou, N.; and Patterson, D. 2001. "XML Query Re-Write Using the Namespace Concept for Information Integration", submitted.
|
|