1 HGIS011006

67 %
33 %
Information about 1 HGIS011006

Published on October 25, 2007

Author: Miguel

Source: authorstream.com

Slide1:  Geographic Information System (GIS) and Public Health 江博煌 國家衛生研究院 衛生政策研發中心 科技管理師 1/10/2006 衛生地理資訊研討系列 3S 科技:  3S 科技 遙測 Remote Sensing (RS) 全球衛星定位 Global Positioning System (GPS) 地理資訊系統 Geographic Information System (GIS) Slide3:  Traditional GIS Design A computerized system for the storage, analysis, and display of geographically referenced information: Data Layers (example) Population Transportation Services Slide7:  Thiessen Polygons Concentric Circles surrounding points have been used in the past for analysis. This has 2 limitations – circles overlap and also leave gaps (see top figure). An improvement over this is the Thiessen Polygon (see bottom figure). In Thiessen polygons, a transect is drawn between two points creating the “best-fit polygon.” Overlap Gap Theissen Polygons Concentric Circles Slide12:  Distance Decay Slide16:  Selected Analytical Methods for Use Public Health Vector Methods Raster Methods Object-Oriented Methods Statistical Methods Slide17:  Vector GIS Source: http://www.geography.hunter.cuny.edu/courses/gis2/lectures/lecture2/vector.gif Slide18:  System Development - Atlas of Health (http://atlas.nhri.org.tw) Slide19:  System Development - Atlas of Health (http://atlas.nhri.org.tw) Slide20:  System Development - Area Resource File (ARF) (http://hgis.nhri.org.tw) Slide21:  CDC 類鼻疽病例分佈 Slide22:  http://www.cdc.gov/hiv/graphics/images/L206/L206-14.htm US CDC Use of GIS Slide23:  US AIDS Cases Animation http://www.cdc.gov/hiv/graphics/images/dotmaps/dotmapan.htm Bird Flu:  Bird Flu Slide25:  Color coded circles for wild bird H5N1:(confirmed or suspected in confirmed location) May = Blue June = Orange July = Yellow August 15 = Green August 31 = Red September = Purple Current = White Green or Red squares are unconfirmed bird deaths in unconfirmed locations Slide26:  Raster GIS Data Model Source: http://www.geography.hunter.cuny.edu/courses/gis2/lectures/lecture2/fig210.gif Raster Methods:  Raster Methods Interpolation Density Raster Interpolation:  Raster Interpolation IDW (Inverse distance weighted) Search Neighborhood Global polynomial Local polynomial Radial basis functions Kriging Raster Interpolations: Kriging:  Raster Interpolations: Kriging Definition: Kriging creates a surface of predicted values using mathematical as well as statistical models. Kriging takes spatial autocorrelation into account Examples: Continuous surface of population at risk Model environmental risk factors such as toxic soil Continuous surface of disease and disease diffusion Slide30:  Spatial Interpolation Method for Identifying Heavy Metal Soil Contamination Areas 行政院環境保護署 國家衛生研究院 環境衛生與職業醫學研究組 國家衛生研究院 衛生政策研發中心 Data:  Data 2002 ChungHua city topsoil monitoring data (1474 topsoil sub-sample) 2004 ChungHua City health survey data (500 human sample – additional 500 in 2005) Slide33:  Sample Statistics and Threshold Limits* (N=1,474) Slide36:  Ordinary Kriging (a) Cd (b) Cr Slide37:  Simple Kriging (a) Cd (b) Cr Slide39:  Spatial Interpolation Method for Identifying 台南安順 Sediment Contamination Areas 行政院環境保護署 國家衛生研究院 環境衛生與職業醫學研究組 國家衛生研究院 衛生政策研發中心 Slide44:  Example: Simple Density Source: http://www.geo.hunter.cuny.edu/%7Emarcofj/gis2/lab3bis/lab3bis.html Other Useful Functions:  Other Useful Functions The CS137 (radioactive isotopes) contamination and thyroid cancer data was supplied courtesy of the International Sakharov Environmental University. Slide46:  Cellular Automata Agent-Based Modeling Object Oriented Methods Slide47:  Cellular Automata: SLEUTH Slope, Land-use, Exclusion, Urban extent, Transportation, Hillshade Urban Sprawl Model using CA developed by Dr. Keith Clarke, University of California – Santa Barbara Uses historical data to simulate present day urbanization Current work: Run the model into the future Simulate alternative futures Compare across scale and cities Apply to Urban Dynamics to cities Source: Keith Clarke, http://www.ncgia.ucsb.edu/projects/gig/v2/About/ppt-pres/gistw.pdf Slide48:  Example: Cellular Automata SLEUTH Model Source: Keith Clarke, http://www.ncgia.ucsb.edu/projects/gig/v2/About/ppt-pres/gistw.pdf Slide49:  Cluster Detection Geographical Analysis Machine (GAM) Spatial Scan Statistic DYCAST Statistical Models Slide50:  Cluster Detection: GAM Geographical Analysis Machine (GAM) Developed by Openshaw, Charlton and Craft (1988)* to assess Childhood Acute Lymphoblastic Leukemia GAM includes a test for statistical significance of clusters Steps: Draws fine grid around study region Draws different-sized circles around the centroid of each grid Determines if incidence inside each circle is higher than expected based on the “typical” rate of illness in the population * Openshaw, S., Charlton, M., & Craft, A. (1988). Searching for Leukemia Clusters Using a Geographic Analysis Machine. Papers of the Regional Science Association. 64: 95-106. Slide51:  Example: GAM Distribution of Significant Acute Lymphoblastic Leukemia Circles * Openshaw, S., Charlton, M., & Craft, A. (1988). Searching for Leukemia Clusters Using a Geographic Analysis Machine. Papers of the Regional Science Association. 64: 95-106. Slide52:  Cluster Detection: SatScan Spatial Scan Statistic (SatScan) – FREE! NOT a “true” GIS as it does not have mapping functions but results can be transferred into a GIS for mapping. Developed by Kulldorf, Feuer, Miller and Freedman (1997)* to test for Breast Cancer Clusters in the Northeast USA. Similar to GAM but uses Monte carlo Simulation to determine if cases are solely due to random chance. Tests include: Poisson – Case clustering based on Population at Risk Bernoulli – Case clustering based on Control group Space-Time Permutation – Case clustering based on Time * Kulldorff M. and Information Management Services, Inc. (2003). SaTScan v4.0: Software for the spatial and space-time scan statistics. Online Internet. http://www.satscan.org/ Slide53:  Dynamic Continuous Area Space-Time (DYCAST) Hunter College, CUNY (2003), to assess areas of high risk of West Nile in NYC. A grid is meshed across NYC while a roving window moves across the grid. At each cell’s centroid, the Knox statistical method checks the probability that dead crows found within the window are statistically close in space (0.25 miles) and time (3 days). Successfully predicted human cases of West Nile in 2001 in NYC. NYC Department of Health used it in 2002 to target areas for spraying pesticide to reduce mosquito population. * Theophilides, C. N., Ahearn, S. C., Grady, S., Merlino, M. (2003). Identifying West Nile Virus Risk Areas: The Dynamic Continuous-Area Space-Time System. Am. J. Epidemiol. 157: 843-854 Slide54:  Source: http://www.geography.hunter.cuny.edu/~yllik/gis2/lectures/lecture6/lecture6.html DYCAST Slide55:  Source: http://www.geography.hunter.cuny.edu/~yllik/gis2/lectures/lecture6/lecture6.html DYCAST Slide56:  Health Informatics R&D (HIRAD) Research Group - Health GIS (HGIS) Research Lab Thank You!

Add a comment

Related presentations