Keywords: spatial statistics, visualization, cartography, public health
Spatial statistics is different from conventional statistics in that attributes for close geographic features are often autocorrelated, which violates the assumption of independence in conventional statistics. The first law of geography states that everything is related to everything else, but near things are more related than distant things (Tobler, 1970). Geographers have been using various methods to measure and visualize spatial autocorrelation for different geographic features. It is not only art but also the science of illustrating geographic features on a map. On the E-poster presentation, I will show the art and science of spatial statistics and visualization. Spatial statistics such as (dual) Kernel Density Estimation, Join Count Statistics, global and local Moran’s I, and Gi statistics will be discussed and illustrated. In addition, applications of spatial statistics in public health will also be presented with a focus on spatial epidemiology that studies the spatial distribution of health outcomes. For example, are residential locations of cancer patients tend to be clustered in space? What are the spatial predictors, such as environmental factors and socioeconomic status, that explain the spatially clustered cancer patients? Spatial statistics will uncover the myths behind these health outcomes. Fortunately, cartography is always the best visualization tool and language for spatial statisticians to tell stories to the audience.