The Division of Biostatistics at the Department of Preventive Medicine, UTHSC, invites you to attend the following seminar.
Time: Monday, July 1, 2:00 pm-3:00 pm CT
Location: 4th Floor Conference Room 400 in the Doctors Office Building at 66 N. Pauline Street, Memphis, TN 38105.
ZOOM Virtual Room Connection: Register in advance for this meeting
Seminar Website: https://www.eventcreate.com/e/biostatisticsseminar
Modeling GeoSpatial Data: Turkish Health Studies
Mehmet Koçak, Ph.D.
Biostatistics and Medical Informatics,
Istanbul Medipol University, Istanbul, Turkey
Spatial autocorrelation is a fundamental concept in spatial statistics, describing the degree to which a set of spatial data points are correlated with each other across a geographical space. This presentation provides a comprehensive overview of spatial autocorrelation, discussing its definitions, detection methods, and modeling techniques. Firstly, we delve into the definition of spatial autocorrelation, highlighting its significance in understanding geographical data patterns. We cover various aspects such as self-correlation due to geographical ordering, information content in geo-referenced data, and its role as a diagnostic tool for spatial model misspecification. Next, we explore methods for detecting spatial autocorrelation, focusing on Moran’s I and Geary’s C statistics, which provide formal tests for spatial dependency. The presentation explains how to compute these statistics and interpret their results, supported by visual examples and practical applications. We then shift to modeling spatial autocorrelation, demonstrating the use of SAS procedure PROC SPATIALREG. The SPATIALREG procedure enables the application of various spatial models, including Spatial Auto-regressive (SAR), Spatial Durbin Model (SDM), Spatial Error Model (SEM), etc. We also provide examples of geospatial modeling using sets of Turkish Health Studies, which are national surveys having more than 20,000 participants, considering spatial autocorrelation in both the response and predictor variables. We discuss the generation of spatial weight matrices and the fitting of spatial models to the data, highlighting key findings and model performance metrics.