Self-organizing map containing mixtures of NO2, O3, PM2.5, and temperature for Utah 2013-2016

This dataset contains a self-organizing map trained on weekly mean and max values for NO2, O3, PM2.5, and temperature for 1 km2 grid cells in Utah from 2013–2016. This is a high-resolution SOM, with 22,750 nodes. The SOM was created in order to link environmental mixtures to pregnancy outcomes for individuals based on the timing of their pregnancy and their residential location. Included in the data are: U-matrix (.umx). Contains the average distance bewteen each node and its neighbors. Feature weights (.wts). A codebook of each node's weights for features used in training. Best-matching units (.bm). Assigns each observation in the spatiotemporal dataset to it's best-fitting node in the SOM surface. See Skupin & Esperbé (2008) for an introduction to high-resolution SOMs and Agarwal & Skupin (2008) for additional guidance on SOMs.