Spatio-Temporal Change of Support Applied to South Dakota Area Deprivation Index Rankings
Presentation Type
Poster
Student
Yes
Abstract
A common issue in spatial and temporal statistical analysis occurs when we want to make inferences about a variable, but the spatial or temporal support of the observed data does not match the desired support. The process of transforming data to the desired support is referred to as change of support (COS). The traditional approach for performing spatial-only COS is to estimate values based on areal proportions. This method is easy to implement but works only for spatial COS and does not provide measures of uncertainty. Furthermore, there is no reliable way to evaluate the performance of this method. In this project, we employed a spatio-temporal change of support (STCOS) model which allows for both spatial and temporal COS on Gaussian data. The model uses a "bottom-up" approach and a Bayesian hierarchical model framework. This methodology can provide model-based estimates, predictions, and associated measures of uncertainty. We present a case study using the national Area Deprivation Index (ADI) rankings for the state of South Dakota. ADI rankings are used to identify areas of socioeconomic disadvantage at the census block group level. The STCOS model is demonstrated by estimating the ADI rankings of ZIP Code Tabulation Areas in South Dakota.
Start Date
2-6-2024 1:00 PM
End Date
2-6-2024 2:00 PM
Spatio-Temporal Change of Support Applied to South Dakota Area Deprivation Index Rankings
Volstorff A
A common issue in spatial and temporal statistical analysis occurs when we want to make inferences about a variable, but the spatial or temporal support of the observed data does not match the desired support. The process of transforming data to the desired support is referred to as change of support (COS). The traditional approach for performing spatial-only COS is to estimate values based on areal proportions. This method is easy to implement but works only for spatial COS and does not provide measures of uncertainty. Furthermore, there is no reliable way to evaluate the performance of this method. In this project, we employed a spatio-temporal change of support (STCOS) model which allows for both spatial and temporal COS on Gaussian data. The model uses a "bottom-up" approach and a Bayesian hierarchical model framework. This methodology can provide model-based estimates, predictions, and associated measures of uncertainty. We present a case study using the national Area Deprivation Index (ADI) rankings for the state of South Dakota. ADI rankings are used to identify areas of socioeconomic disadvantage at the census block group level. The STCOS model is demonstrated by estimating the ADI rankings of ZIP Code Tabulation Areas in South Dakota.