Prototype Geographic Search and Query Tools for a Library Digital Archive
Continued...

The user clicks the + box beneath gazetteer to expand a hierarchical drop-down list of gazetteer feature types. A gazetteer-based search is performed by clicking any of the check boxes adjacent to the feature type name, such as a construction material and supply business, and clicking the red Search button. All entries and selections in the text-based search tools can be cleared by clicking the red Reset button.

Geographic features are described by three elements: type, name, and geographic footprint. Each feature element is important to the user as well as the Digital Library staff. For instance, a user may be interested in locating library holdings in relation to a named geographic feature or type of feature but not know the exact spatial coordinates of the feature of interest.

The gazetteer can transform a named geographic feature into a spatial representation that is then used to search the database. In addition, in the future the Digital Library staff will be able to use the gazetteer as a tool to specify geographic footprints for library holdings quickly and cheaply. This is important because the geographic search and query tools discussed will work as planned only for Digital Archive holdings that incorporate geographic references.

Summary

The GeoBrowser application was built to provide the USC Digital Archive Web site with advanced spatial search capabilities and includes an interactive map, a geocoder, and a gazetteer. The spatial services include an ArcIMS map, an address geocoder, and a gazetteer. The map allows Web site users to easily define a geographic area of interest on a map, view available holdings within the selected area, and search the Digital Archive for information on those holdings. The ArcIMS map browser was coded in Java, JSF, JavaScript, and CSS. The geocoder is written in C# and hosted as a Web service. The geocoder provides another very popular kind of search tool. The user can enter any address and ZIP Code in the United States to locate library holdings in proximity to the address. The gazetteer affords a lengthy, detailed geographic feature type hierarchy. Gazetteer features are organized by type, name, and geographic footprint, and this information is used as spatial input for searching the Digital Archive. The gazetteer is stored in a MySQL database and accessed via custom JavaScript and Java servlets. All three of these spatial search tools can be used independently to search for library holdings.

References—In Print

Bakshi, R., C. A. Knoblock, S. Thakkar, 2004. "Exploiting Online Sources to Accurately Geocode Addresses," Proceedings of the 12th International Symposium on Advances in Geographic Information Systems, ACM-GIS 2004.

Goldberg, D. W., J. P. Wilson, and C. A. Knoblock, 2005. "Extracting Geographic Features from the Internet to Automatically Build Detailed Regional Gazetteers." Unpublished paper presented at the Second Annual ECAI Cultural Congress, Shanghai, China, May 9�13, 2005.

Goldberg, D. W., J. P. Wilson, and C. A. Knoblock, 2005. "Extracting Geographic Features from the Internet to Automatically Build Detailed Regional Gazetteers." International Journal of Geographical Information Science (submitted 12/2005).

Goldberg, D. W., J. P. Wilson, and C. A. Knoblock, 2006. "Beyond Linear Interpolation: Moving Towards Mediator-Based Geocoding Using Multiple Sources." Journal of the Urban and Regional Information Systems Association (submitted 1/2006).

Wilson, J. P., C. S. Lam, and D. A. Holmes-Wong, 2004. "A New Method for the Specification of Geographic Footprints in Digital Gazetteers." Cartography and Geographic Information Science, 31, 195�207.

References—Online

USC GIS Research Laboratory, "USC Digital Library Archive Prototype Geographic Search and Query Tools." Project application demonstration Web site hosted since 8/2005. (gislab07.usc.edu:8080/ cispubsearch1/AdvancedSearch1.jsf, accessed 1/2006).

Kafle, S., "ArcIMS Java Server Faces (JSF) Web Application," Esri Knowledge Base, (arcscripts.esri.com/details.asp?dbid=14085, accessed 1/2006).

About the Authors

Dr. Jennifer Swift is a research assistant professor in the Department of Civil Engineering at USC who has managed a wide variety of GIS research and Internet map service/data portal development projects. Many of these projects have involved the development of data models, database structures, and custom software modules for geologic, geophysical, and geotechnical engineering applications. She has also participated in the development and deployment of data models and schemas for geotechnical engineering, geophysics, and centrifuge testing. She can be reached at jswift@usc.edu.

Parisa Ghaemi is a second year master's student in the Department of Computer Science and graduate research assistant in the USC GIS Research Laboratory. She has written ArcIMS-based software applications for several USC GIS Research Laboratory projects and plans to pursue her doctorate while focusing her research on ArcIMS application development. She can be contacted at ghaemi@usc.edu.

Daniel Goldberg is a second year doctoral student in the Department of Computer Science at USC and graduate research assistant in both the USC GIS Research Laboratory and Information Sciences Institute. He currently focuses his research on geospatial data extraction and integration. Recent projects have led him to develop new techniques for address geocoding and automatic gazetteer generation that utilize geographic data extracted from freely available online data sources. He can be reached at daniel.goldberg@usc.edu.

Dr. John Wilson, professor in the Department of Geography at USC and director of the GIS Research Laboratory, currently leads numerous research projects and has published two edited books and approximately 60 journals, articles, and book chapters on GIS, spatial analysis, and environmental modeling topics. Many of his projects combine basic and applied research to develop and/or conduct analyses using state-of-the-art geographic datasets. In addition, he has taught numerous Esri-specific workshops and has received numerous awards for his teaching and mentoring of students. He can be reached at jpwilson@usc.edu.