Geographic Information Science - 東京大学arikawa/s-it/3... · Geographic Information Science...
Transcript of Geographic Information Science - 東京大学arikawa/s-it/3... · Geographic Information Science...
Geographic Information Science
David M. MarkProfessor of Geography &
Director, National Center for Geographic Information and Analysis
University at Buffalo
Part 1: Geographic Information
Science and GIS Research at Buffalo
Geographic Information
• Many social, economic, and environmental processes operate at geographic scales, over geographic space
• Thus Geographic Information is relevant to many human activities and decisions
Geographic Information Systems
• In the 1960s and 1970s, software systems were developed to manage geographic information
• In the 1980s, this developed into a commercial GIS software industry
• The University at Buffalo (SUNY) has long been a leading producer of trained professionals in the GIS field
Geographic Information Systems at the University at Buffalo
• In 1975, Buffalo established the first “GIS Laboratory” of that name in a U.S. university
• In 1988, Buffalo and its partners were awarded the National Center for Geographic Information and Analysis by the National Science Foundation
• NCGIA continues both as a national research consortium, and as an Organized Research Unit at Buffalo
• Crime Analysis• Emergency Preparedness• Public Health and Human
Services• Urban and Regional
Planning• Water Resources• Transportation Planning
and Monitoring• Etc.
Major GIS Application Areas
(URISA Journal v 12 No 2 Spring 2000)
Geographic Information Science
• Geographic information science is the information science behind GIS software!
• GIScience develops and augments the theory upon which GIS of the future can be built
• GIScience extends the capability of GIS• GIScience studies the relationships
between GIS and society
Geographic Information Science
• GIScience is sometimes referred to as GeoInformatics
• GIScience emerged around 1990 as a response to gaps in GIS software and technology
Research Projects at SUNY Buffalo Supported by Research Grants
• In the 2000/2001 year, NCGIA managed 20 external grants
• These grants involved 20 different faculty as PI or co-PI
• Those faculty come from 8 different academic departments in 3 Schools (CAS, SEAS, Medical)– Civil, Structural, & Environmental Engineering,
Computer Science & Engineering, Geography, Geology, Industrial Engineering, Philosophy, Political Science, Social and Preventive Medicine
2000/2001 Funding: Other Facts
• Project direct cost budgets currently exceed $1 million per year
• In academic year 2000-2001, NCGIA grants employed approximately 40 graduate students
2000/2001 Funding: Sources
• National Science Foundation (8)• National Institutes of Health (2)• National Imagery and Mapping Agency (2)• CUBRC (3)
– Of these, 2 are US DoT, 1 NYS DoT)
• US Environmental Protection Agency (1)• Others:
– NASA (1), NIJ (1), CDC, USARMC (1), FGDC/UCGIS (1)
NCGIA Members• Computer Science &
Engineering– Stuart C Shapiro– Aidong Zhang
• Economics– Alex Anas
• Geology– Marcus Bursik– Michael Sheridan
• Anthropology– Ezra B Zubrow
• Biological Sciences– Guiyun Yan
• Civil, Structural, & Environmental Engineering– Joseph F Atkinson– Alan Rabideau
NCGIA Members• Geography
– S. Bagchi-Sen– Ling Bian– Hugh W Calkins– Meghan S Cope– Douglas Flewelling– Chris S Larsen– Barry Lentnek– David M Mark– Peter Rogerson– Jean-Claude Thill– Michael J Woldenberg
• Industrial Engineering– Rajan Batta– Christopher M Rump
• Law– Robert I Reis
• Information Studies– Corinne Jorgensen
NCGIA Members• Linguistics
– David A Zubin• Philosophy
– Robert Casati– Barry Smith
• Planning– Hugh S Cole– G. William Page
• Political Science– Munroe Eagles
• Regional Community Policing Center– Pamela K Beal
• Social and Preventive Medicine– Jo Freudenheim– John Vena
• Sociology– Stephen Hart
Some Recent Externally-Funded Projects
Geospatial Lifelines• National Institute of Environmental Health
Sciences, National Institutes of Health• Spatio-Temporal GIS Analysis for Environmental Health• D. Mark, L. Bian, P. Rogerson, J. Vena
with Max Egenhofer, U. Maine*• Geography and Social & Preventive Medicine• $1,563,536 over 4 years*
* includes subaward to the University of Maine
Geospatial Lifelines
• Hot spots and clusters• Influences of human residential mobility
2000
1970
1940
1910
A BC
D
A B C
D
Fred
MaryJohn
Breast Cancer Risk
• National Institutes for Health• Breast Cancer Risk:
Residential environment & genetics• J. Freudenheim• Social & Preventive Medicine• $231,000 over 2 years
Breast Cancer Research
• USARMC, US Army• Environmental Exposures at Birth and Menarche and
Risk of Breast Cancer• J. Freudenheim and D. Mark• Social & Preventive Medicine and Geography• $153,476 over 3 years
Angler Cohort Study
• CDC/ATDSR (Center for Disease Control)• The New York State Angler Cohort Study:
Reproductive and Developmental Health• J. Vena• Social and Preventive Medicine• $4,339,908 over 3 years
Geographical Changes in Crime
• National Institute of Justice• Detection and Prediction of Geographical
Changes in Crime Rates• P. Rogerson, R. Batta, and C. Rump • Geography and Industrial Engineering• $221,520 over 2 years
Crime Analysis
A Socio-Economic Model of theGeographical Displacement of Crime
• Model to predict the number of crime incidents within a police jurisdiction
• Expected reward is a function of wealth, crime level & police enforcement
• Goal is to determine the best allocation of police enforcement resources
n*
Opportunity Cost of crime
n(1) n(2)
mE[R]
Crime Level
Monitoring Spatial Patterns
• National Science Foundation, Geography and Regional Science
• Methods for the Monitoring of Spatial Patterns• P. Rogerson• Geography• $93,763 over 2 years
Quick Detection of Geographic Clusters of Crime
• Methods of spatial analysis that are used in finding clusters of crime are often retrospective
• Alternative: Repeated testing of patterns of crime as new data become available
• Objective is to uncover new geographic patterns quickly
1996 Arsons
Blue Dots Represent Successive Arsons Leading to New Geographic Cluster Signal
Immersive Information Spaces• National Imagery and Mapping Agency• Immersive Information Spaces for Data Access and
Dissemination• S. Fabrikant, UCSB, D. Mark, UB,
and . D. Montello, UCSB
$366,071 over 3 years*
Image DatabasesThis project focuses on the automatic identification and This project focuses on the automatic identification and
retrieval of industrial building sites from digital aerial retrieval of industrial building sites from digital aerial photographsphotographs
Sample Image
Retrieved Images of industrial sites in Buffalo
Digital Government• National Science Foundation, Digital
Government Program• Very Large Scale Multidimensional Data Management
and Retrieval for USGS and NIMA Imagery• A. Zhang and D. Mark• Computer Science & Engineering and
Geography• $500,000 over 3 years
Digital Libraries, Japan Collaboration
• National Science Foundation, Digital Libraries Program
• Metadata Model, Resource Discovery & Querying on Large-Scale Multidimensional Datasets
• A. Zhang and D. Mark at Buffalo• Frederic Andres at National Institute of
Informatics• $400,000 over 3 years
Satellite Imagery Analysis
• National Aeronautics & Space Administration• Tephrostratigraphic and satellite-imagery analysis of
Kamchatka shoreline: Great earthquakes and mega-eruptions of Kamchatka, Russia
• M. Bursik and L. Bian• Geology and Geography• $110,000 over 2 years
Multi-Resolution Data Retrieval
• National Imagery and Mapping Agency • A Multi-Resolution Geographical Data Retrieval System • L. Bian and A. Zhang• Geography and Computer Science &
Engineering• $75,125 over 3.3 years
GIS Calibration Tools for Groundwater Modeling
• U.S. Environmental Protection Agency• GIS-Based Calibration Tools for a High Performance
Analytic Element Model• A. Rabideau, M. Becker, and D. Flewelling, • CSEE, Geology and Geography• $997,100 over 3 years
GIS Calibration Tools for Groundwater Modeling
125+ tons of hazardous chemicals per year
No comprehensive model of groundwater flow in Niagara County... so it could be more
Creating a framework for building regional models of groundwater flow from geographic data
Project funded by US EPA
SpatialDB
Engineering Models
GIS User Interface
26 High LevelHazardous Waste
Sites in Niagara andErie Counties
Transportation Improvements
• NYS Department of Transportation (via CUBRC)
• Economic Development & Transportation Improvements• J-C Thill• Geography• $85,636 over 1.25 years
Emergency Response: Transportation
• Highway fatalities and air medical transport services
• Uses Fatal Accident Response System (FARS) data
• Collaborative work with Veridian
Urban and Regional Planning, Regional Economic Development• REDD- Regional Economic
Development Database• UB Institute for Local Governance
and Regional Growth
Ontology
• National Science Foundation, Geography and Regional Science
• Geographic Categories: An Ontological Investigation• D. Mark and B. Smith• Geography and Philosophy• $199,689 over 3 years
Mountains are the Best Example of Geographic Things
•
QuickTime™ and aPlanar RGB decompressor
are needed to see this picture.
Integrative Graduate Education and Research Training
• National Science Foundation, Division of Graduate Education
• Integrative Graduate Education & Research Training in Geographic Information Science
• D. Mark, R. Batta, B. Smith, E. Zubrow • 8 academic departments• $2,349,350 over 5 years
IGERT in GIScience• Total funding $2.3M over 5 years• Buffalo’s GIScience IGERT currently
supports 21 Ph.D. students in four departments
Part 2: An Overview of
Geographic Information Science
Outline of Part 2
• Introduction• Components of Geographic Information
Science• History of the Field• Summary and Prospects
Geographic Information Science
• "Geographic Information Science (GIScience) is the basic research field that seeks to redefine geographic concepts and their use in the context of geographic information systems.”
• “GIScience also examines the impacts of GIS on individuals and society, and the influences of society on GIS.”
Geographic Information Science
• “GIScience re-examines some of the most fundamental themes in traditional spatially oriented fields such as geography, cartography, and geodesy, while incorporating more recent developments in cognitive and information science”
Geographic Information Science
• “It also overlaps with and draws from more specialized research fields such as computer science, statistics, mathematics, and psychology, and contributes to progress in those fields." (NSF Workshop)
What GIScience is Not
• Geographic Information Science is not just a new name for GIS applications!
• Even the use of GIS as a tool in scientific research is not Geographic Information Science— at a recent NSF workshop, this latter area was termed "Research Using GIS”: – Research Using GIS, or RuGIS, is important to the sciences,
and to funding for GIS-related scientific work, but is not GIScience per se
Components of GI Science
• The nature of a scientific field can be characterized by listing its components
• This is not of course the first attempt to list components for GIscience, and earlier efforts will be reviewed later
• This order was chosen so that the new proposed structure could be used to frame the review of previous efforts
Components of GI Science
• These components are not meant to define a cutting-edge research agenda
• It is more like a curriculum for the field, the basic subfields, even if research on some of them is relatively mature
• Also, these headings are neither definitive nor exhaustive, and several important topics that do not readily fit into the scheme are reported under the heading “Loose Ends” below
1. Ontology of the Geographic Domain
• Ontology deals with what exists, and with what may possibly exist
• In this sense, it is a branch of philosophy that deals with some of the most fundamental aspects of scientific inquiry, but at a very high level of abstraction.
• Ontology seeks to provide a consistent formal theory of tokens (instances) and types (kinds) in the real world, their relationships, and the processes that modify them
2. Formal Representation of Geographic Phenomena
• More recently, the term ontology has been used in information science and knowledge representation to refer to the specifications of the conceptualizations employed by different groups of users in regard to domains of objects of different types
• The resulting representations are types in the digital domain, to be instantiated through data to become digital tokens (data instances) that correspond to geographic things in reality
3. Acquisition and Quality of Geographic Data
• Data (measurements) of positions and attributes of the geographic domain is still central to GIS
• Technologies for the acquisition of geographic data and information, chiefly remote sensing and GPS, are important here
• The accuracy of geographic information, also termed data quality research, is included here but could be listed as a major heading in its own right
4. Qualitative Spatial Reasoning
• Reasoning about spatial relations and positions is a well-established research area in artificial intelligence, and has become important in GIScience as well.
• There are basically two domains for spatial relations:– Distance and direction for disjoint spatial objects – (Mereo)Topological spatial relations for spatial objects that
touch or overlap (9 intersection or RCC)• Relations can be characterized on quantitative or
qualitative scales
5. Computational Geometry
• Computational geometry provides fundamentals for metric representation of objects and relations in geographic space
• Efficient computation of proximity, handled under the conceptual framework variously labeled as Voronoidiagrams, or Thiessen or proximal polygons, is especially important
• Analytical cartography is an alternative term for many aspects of computational geometry applied to the geospatial domain
• Map ‘generalization’ also falls here
6. Efficient Indexing, Retrieval, and Search in Geographic Databases
• Efficient indexing of multidimensional data is an important problem in database research in computer science.
• Since geographic information is inherently at least two dimensional, these indexing issues have long been important in GIS. – Morton 'matrix' approach for ordering map areas in CGIS
(Morton, 1966). – Quadtrees (see Samet, 1989, for a review)– Many related indexing schemes such as B-trees, R-trees, k-
d trees, etc.
7. Spatial Statistics
• Spatial statistics is an important research area with strong links to Geographic Information Science.
• Spatial autocorrelation or dependence is often present in geospatial information
• Spatial statistics (Cressie, 1993) provides formal statistical methods for characterizing and dealing with spatial autocorrelation
8. Cognitive Models of Geographic Phenomena
• This research area involves the study of human perception, learning, memory, reasoning, and communication of and about geographic phenomena
• An explicit examination of human cognition of geographic things was originally introduced into the GIScienceagenda as a way to gain insight into geographic ontology, spatial relations, and human-computer interaction for GIS (Mark and Frank, 1991)
• The COSIT meetings attest to the strength and breadth of this research component
9. Human Interaction with Geographic Information and Technology
• Human-computer interaction (HCI) for geographic information systems, and the design of user interfaces for GIS, are perhaps the most obvious example of the relevance of cognition to GIS
• If geographic concepts and HCI cannot be dealt with separately, then the GIScienceresearch community must address problems in the overlap
10. Geographic Information, Institutions, and Society
• Societal impacts of GIS technology have been a studied within GIS since its onset
• Indeed, before 1980, most technical innovation in GIS occurred in a direct application context
• During the 1980s, commercial software matured and an academic GIS R&D agenda developed
• In the 1990s, there was an increase in research on economic and legal aspects of geographic information, plus a post-Modern critique of GIS
• These trends continue
(11. Spatial Analysis?)
• Is the development of new methods of spatial analysis part of GIScience?
• Or does spatial analysis represent an application of the results of GIScience to problems in environmental or social science?
• Is “spatial analysis” a useful grouping of GISciencetopics? Or is it already covered by the ontology, representation, and spatial statistics topics?
• MAUP seems clearly to be a GIScience issue, even if spatial analysis in general is not
(12. Missing Topics?)
• Where does map algebra (Tomlin 1990) fit into this conceptual scheme?
• What about scale?• What about time?• What about GeoVisualization?• What about GeoComputation?• What other topics in GIScience are
missing?
Other Lists of Components• Mike Goodchild made a keynote address about GIS-
related research priorities to the Spatial Data Handling conference in Zurich, Switzerland in 1990, entitling the talk “Spatial Information Science”
• The address was subsequently published in 1992 in the International Journal of Geographical Information Systems under the modified title “Geographical Information Science”
Goodchild’s definition
• In the IJGIS article, Goodchild did not provide a crisp definition of the field, but did discuss what he called "the content of geographical information science", under the following eight headings:
Goodchild’s “Content of geographical information science”
• 1. Data collection and measurement• 2. Data capture• 3. Spatial statistics• 4. Data modeling and theories of spatial data• 5. Data structures, algorithms and processes• 6. Display• 7. Analytical tools• 8. Institutional, managerial and ethical issues
The University Consortium for Geographic Information S...
• A series of discussions through 1992 and 1993 led to a founding meeting in Boulder, Colorado, in December, 1994
• Several words in the name were debated, and votes were taken
• The delegates voted to make the name the "University Consortium for Geographic Information Science", rather than UCGI Systems or UCGI & Analysis.
• In June 1996, 10 UCGIS research priorities were identified.
UCGIS Research Challenges in GIScience (1996)
• Spatial Data Acquisition and Integration• Distributed Computing• Extensions to Geographic Representation• Cognition of Geographic Information• Interoperability of Geographic Information• Scale• Spatial Analysis in a GIS Environment• The Future of the Spatial Information Infrastructure• Uncertainty in Spatial Data and GIS-based Analyses• GIS and Society
UCGIS Emerging Research Themes (2001)
• Ontological Foundations for Geographic Information Science
• GeoVisualization• Geographic Data Mining & Knowledge
Discovery• Remotely-Acquired Data & Information
Comparison of Components
All Three Agree
GIS and SocietyInstitutional, managerial and ethical issues
Geographic Information, Institutions, and Society
GI & Society
Spatial Data Acquisition and Integration; Uncertainty in Spatial Data and GIS-based Analyses
Data collection and measurement; Data capture
Acquisition and Quality of Geographic Data
Acquisition
Extensions to Geographic Representation
Data modeling and theories of spatial data
Formal Representation of Geographic Phenomena
RepresentationUCGISGoodchild 1992DMM schemeKeyword
Two of Three Agree
-Spatial StatisticsSpatial StatisticsStatistics
-Data Structures, Algorithms, and Processes
Efficient Indexing in Geographic Databases
Indexing
-Data Structures, Algorithms, and Processes
Computational Geometry
Geometry
Cognition of Geographic Information
-Cognitive Models of Geographic Phenomena
Cognition
Spatial Analysis in a GIS Environment
Analytical Tools(Spatial Analysis)AnalysisUCGISGoodchild 1992DMM schemeKeyword
No Agreement• Goodchild
– Display• UCGIS
– Distributed Computing– Interoperability of Geographic Information– Scale– The Future of the Spatial Information Infrastructure
• This Paper– Ontology of the Geographic Domain– Qualitative Spatial Reasoning– Human Interaction with GI and Technology
Summary and Prospects• Geographic Information Science has
considerable depth, and a richness of intellectual challenges that mark it as a legitimate multidisciplinary field and perhaps an emerging new discipline
• It is my opinion that progress in the field would be aided by a consensus among leading researchers on the nature, scope, and elements of the field
Contacts:
• Thank you for your attention!
• David M Mark, Director, NCGIA– [email protected]
• Web: http://www.geog.buffalo.edu/ncgia/