Software

This package provides functipons for fitting a geostatistical downscalers to integrate monitoring data with a proxy (e.g., satellite imagery or numerical model simulations). The package also includes functions to fit spatial ensemble to combine predictions and to conduct various cross-validation experiments for performance evaluation. This work was originally created for supporting the NASA MAIA mission. 

Publication: Remote Sensing | 2025 | Madden et al. 

Maintainer: Wyatt Madden

The goal of clbart is to estimate heterogeneous effects of a primary exposure in a scenario where only cases are observed. Specifically, this method is appropriate for data which have been constructed using the design, such that only one case exists within each strata. The effect of the primary exposure is modeled as a function of time-invariant covariates using Bayesian additive regression trees.

Publication: JASA | 2025 | Englert et al. 

Maintainer: Jacob Englert

This codebase includes the function and tutorial for fitting double GLM model that allows for time-varying covariate-dependent dispersion parameters. This work is motivated by the time-series analysis of count data and environmental exposure, where improved characterization of dispersion can lead to better precision for the exposure risk estimates. 

Publication: Epidemiology | 2025 | Zhang et al. 

Maintainer: Danlu Zhang

This package provides functions for fitting a negative-binomial regression model that conceptualizes exposure quantile functions as the exposure of interest. The use of a functional (or distributional) covariate is to account for exposure heterogeneity for studying short-term effects of environmental exposures across time and/or space when the outcome is aggregated counts. 

Publication: Biometrics | 2024 | Zhang et al. 

Maintainer: Yuzi Zhang

This package provides functions for fitting Bayesian time-series or spatial-temporal negative-binomial (NB) model to  with time-varying coefficients that incorporates measures of infection activity. Tools to estimate attributiable deaths by comparing observed counts versus a counterfactual of no infection activity are also provided.  

Publication: JRSS A | 2025 | Zhang et al. 

Maintainer: Yuzi Zhang