Bayesian Methods for Socio-Spatial Point Patterns and Networks

Ohio State University, The, Columbus OH

Investigators

Abstract

This project seeks to make contributions to statistical science in the areas of spatial statistics, network analysis, and Bayesian parametric and nonparametric modeling. Underlying the methodological aspects of the research is a particular area of application, contextual effects and exposure analysis research. Statistical methodology will be developed to improve the ability to characterize human activity patterns and social networks using state-of-the-art large-scale sample survey data. Specific contributions include the development of novel Bayesian stochastic models for multivariate spatial point patterns based on dependent nonparametric latent intensity functions. Statistical models will also be developed for spatially-referenced social networks; these models describe relational ties based on proximity in a random deformation of geographic space. Both classes of models will be thoroughly studied from both a theoretical and empirical perspective. The statistical research focus of this project is deeply rooted in an area of application bridging the social, geographic, and health sciences. Research on contextual effects aims to understand the consequences of social and environmental exposures on individual outcomes. Efforts to quantify the influence of the multiple contexts to which individuals are exposed are limited due the inherent complexities of existing data sources, as well as those currently being collected using state-of-the-art GPS-based technologies. This project aims to overcome these data challenges by introducing novel statistical methods and is expected to contribute generally to the field of statistics, as well as the motivating area of application. Educational and training components of the project will reflect this multidisciplinary research setting.