With increasingly abundant spatial data in the form of case counts or rates combined over areal regions (e.g ZIP codes, census-tracts or counties), interest turns to formal identification of difference "boundaries", or barriers on the map, rather than the estimated statistical map itself. "Boundary" refers to a border that describes vastly disparate outcomes in the adjacent areal units, perhaps caused due to latent risk factors. This article focuses on developing a model based statistical tool equipped to identify difference boundaries in maps with a small number of areal units, also referred to as small scale maps, where detecting "boundaries" becomes relatively more challenging. This article proposes a novel and robust nonparametric boundary detection technique, coined as the Dirichlet ProcessWombling (DPW) rule, by employing Dirichlet Process based mixture models for small scale maps. Unlike the recently proposed nonparametric boundary detection rules based on false discovery rates, DPW is free of ad-hoc parameters, computationally simple and readily implementable in freely available software for public health practitioners such as JAGS and OpenBUGS and yet provides statistically interpretable boundary detection in small scale wombling.

1. "connecticut_hypo_accept.txt": The set of all hypotheses from Dirichlet Process which are compatible with the Connecticut map
2. "sbn1.txt": the BUGS code to run the model with continuous rate response in different counties
3. "sbnpois.txt": the BUGS code to run the model with count response in different counties 4: "jags_script.R": Calls BUGS code within R through rjags package, calculates posterior probability of detecting statistically significant boundaries between two geographical units in the Connecticut map and compares the results with CAR model. The code only includes the implementation for the continuous response normal model. Post processing can be done similarly for the count response Poisson model.