cooper

Document Sample
cooper Powered By Docstoc
					Position Paper by Colin A. Cooper and Dawn J. Wright, Dept. of Geosciences,
Oregon State University
For a session titled Mapping Humanity’s Knowledge and Expertise in the Digital
Domain, to be held at the Annual Meeting of the Association of American Geographers
(AAG), Denver, CO, April 5-9, 2005, organized by Katy Börner & André Skupin


Part One
                                         Main Interest
        The results of mapping humanity’s knowledge and expertise in the digital domain are
of extreme importance for enhancing and extending a global communication. This
communication is important for bringing data together to streamline information.
Spatialization of knowledge is a key way for people to highlight similar ideas and build a
conversation with experts in particular areas. Using a metaphorical map framework
establishes a sense of place for non-spatial information. Users of this implemented
technique will not only be able to find what they are looking for, but also see relating fields.
For this holistic endeavor to happen, information visualizations must be constructed
effectively, allowing users to search efficiently while understanding the overall scope. This is
particularly important now in the marine and coastal resource community, as users are
seeking disparate data sets, internal documents, undocumented data sets, and published
reports to aid them in making decisions on management, protection, and conservation in a
dynamically changing marine environment. We have not seen anything in the current
literature with a focus on spatialization of this kind of knowledge for these purposes. As
such, this project aims to spatialize a coastal reference database critical to ongoing ocean
policy initiatives in Oregon (e.g., www.oregonocean.org), thus increasing usability.
                                    Technical Challenges
        The most difficult challenge that we currently see is mapping everything in a sensible
way that can be searched with ease, allowing instant availability of the desired information,
providing a way that can help the user make decisions as to where to limit or increase access
to a coastal zone, which regions are hazardous to recreational activities, which species are
most resilient to changing conditions over months to years (and should they therefore be
protected?), where gaps exist in data coverage, etc. While the Internet is the best medium to
achieve this goal, an effective map metaphor must be created using sound cartographic
principles. Label rendering will be of utmost importance in searching the map. When
delineating searchable areas, borders will also be important, but not enough to enable fast
searching. Landscape/seascape creation and use of symbols and colors will be critical to
give the users a sense that they are looking a diverse, differentiated landscape. Will
traditional topographic and political map templates be enough to accomplish a map
metaphor that is both contemporary and user friendly, in addition to being useful for the
needs of a specific community of our primary users: coastal resource planners, managers,
and marine ecologists? As such, we anticipate the need for graphic illustrations on the map
to help orient the user, similar to illustrative tourist maps. Using GIS products such as
ESRI’s ArcMap and ArcIMS, interactive maps can be created that allow various search
methods and the ability to add graphic symbols, but a final challenge is the implementation
of a seamless and reliable interface between ArcIMS and our Access database of non-
geographic data via an Active Server Pages connection.
                                 Non-Technical Challenges
        There are several challenges to mapping out humanity’s knowledge and expertise,
again in the context of the marine/coastal application domain. Once a user has access to a
search an all encompassing map, will they know how to use it? Traditional cartographic
principles will have to be implemented successfully so that the user does not get lost in a
world of words. Above that, the everyday non-map user has to become familiar with using
interactive maps efficiently.
        A paramount challenge to creating knowledge maps and making them available to
everyone is security. Making knowledge and expertise available entails responsibility. There
is potential for a group of users not being able to handle the information appropriately, such
as a coastal conservation citizen’s group, a K-12 class doing a coastal project, or even
planners or managers who may be unfamiliar with maps, let alone knowledge maps. This
also brings up questions of willing participation of leading experts. What information (e.g.,
occasionally controversial fisheries logbook data) can be made available freely and what must
be restricted (copyrighting or private). Finally, it is a massive undertaking to ensure that
everyone and everything is included with minimal bias. With such a large endeavor,
standards in (metaphoric) map making should be developed to ensure continuity.
                                Opportunities Envisioned
       Our overall goal is to add the recently obtained non-geographic coastal/marine
information to our existing Oregon Coastal Atlas (www.coastalatlas.net), an
infrastructure for data sharing, spatial analysis, resource decision-making, and ocean
policy-making . In general, making information more accessible and available to people
will increase global communication. Students and researchers who think they are alone
will be able to find others interested in the same or related fields. Knowledge and
expertise maps can increase coordination and reduce redundancy. Finally, the crux of
mapping out humanity’s knowledge and expertise is the hope that a greater pattern will
be observed, ultimately leading to new knowledge construction and insight.
Part Two
                      Project Name, Members and Web Address
       The research in progress involves building visualizations for the Catalogue of
Oregon Marine and Coastal Information (COMCI), as a new component of the
aforementioned Oregon Coastal Atlas, using spatialization schemes. Members of the
broader Oregon Coastal Atlas include the authors, (geography master’s degree candidate
Colin Cooper and Professor. Dawn Wright at Oregon State University), as well as grad
students P. Bower, C. Zanger, and D. Pattison of Oregon State, T. Haddad, P. Klarin, and
R. Dana of the Oregon Ocean-Coastal Management Program, and M.Dailey of the non-
profit NGO Ecotrust. The current web address of COMCI is
http://terrene.science.oregonstate.edu/website/comci/comci.asp.
                                  Targeted User Group
       The COMCI is a database of references (published and unpublished, internal
government documents, books, legal documents, etc.) pertaining to information on
Oregon’s marine environment. Its purpose is to provide fast and efficient access to
important information for coastal resource managers, policy makers, researchers, and
interested members of the public. Part of creating a successful policy for managing
ocean and coastal resources involves use of the best available science. COMCI not only
offers access to the best available science, but also incorporates information on coastal
zone management in general and includes links to organizations involved in management.
                                  Supported User Tasks
         Currently, COMCI allows users to search by keyword, author, location, topic
group and major theme. An on-line interactive map was created using ArcIMS to search
the COMCI database. Most records in the database have been assigned to the
geographical areas to which they pertain. This allows the user to focus in on his/her area
of interest while, at the same time, being able to see the surrounding and related fields.
The map fails to portray the records that do not specifically address geographic areas.
Using spatialization techniques, all the records can be included and displayed using a
map metaphor. This will allow users an innovative new way to access information and
gain new knowledge by tapping into human’s inherent ability for spatial reasoning (e.g.,
Fabrikant, 2000).
                                        Data Sets Used
         The COMCI database currently exists as an Access database containing
information on hundreds of records referencing marine and coastal information. Each
reference contains a short descriptive summary, citation, theme, and other ancillary
information. Specific locations are matched to pre-defined areas of the Oregon Coastal
Atlas.
                              Proposed Methods and Maps
         To create an information space, each record in the database must be given a
location based on its similarity to other records. Two techniques will be analyzed: multi-
dimensional scaling (MDS) and the Self Organizing Map (SOM). The most appropriate
technique will be chosen depending on availability of software and hardware. The SOM
is an attractive technique due to its availability through GeoVista, an open source
software package. Also, depending on the data quantity and number of dimensions, the
SOM technique can achieve more accurate spatializations than MDS due to its pre-
defined space (e.g., Jongh and Ormeling, 2003).
         Once an information space is created using spatialization techniques, the map
metaphor will be used to visualize the dataset. Records in the COMCI database have
associated publication dates. If possible, a time animation will be created showing the
information that populates the space.
De Jongh, C. and Omerling, F. (2003). Mapping Non-Spatial Phenomena. Proceedings of
   the Seminars on Developing the ICA-CET Internet Cartography Course.

Fabrikant, S. I. (2000). Spatialized Browsing in Large Data Archives. Transactions in
   GIS, vol. 4, no. 1: 65-78.