elwood-spatial: Information-Theoretic Outlier Detection for Spatial Networks

Much of this code has been ported over from operational scripts currently doing outlier detection for production feeds. Everything was specific and procedural, and I have made efforts to abstract out the core logic, and make this less tied to surface observations of air quality, and make it broadly applicable to any graph/network structure — although it's primarily designed for spatial, environmental monitoring networks. I provided a pattern and logic and co-authored this with Claude. As this is in beta, there is a lot that can change, particularly with the scope and functionality of this package.

This draft release will be used to generate a DOI for a paper that references this method, hopefully to be published within the year. Please contact with questions.