This page lists software and tools I have built, as well as one-off code functions that I think could be useful.


Macro for Compilation of Report-backs (MCR)

This repository contains the code for the Macro for Compilation of Report-backs (MCR), an open-source, generalizable tool designed to provide researchers in health studies with a flexible and scalable method for compiling individual reports. If you use MCR in your projects, please cite Polka et. al., 2021. Let me know if you have any questions!


Framework for Rapid Exposure and Health Effects Estimation (FRESH-Est)

The Framework for Rapid Exposure and Health Effects Estimation (FRESH-Est) is a tool that allows users to estimate the health affects attributable to point source emissions. The tool combines dispersion modeling (AERMOD) with calculations of attributable burden (similar to the EPA’s BenMAP program) in a single framework. The benefits of using FRESH-Est are: 1) health effects for various emissions scenarios can be estimated quickly, and 2) there is an optimization module that focuses on receptor concentrations, health effects, or inequality impacts. See the Milando, Martenies, and Batterman, 2016 for more details. A working demo is also available in this google folder (access available upon request).


Two-stage synthetic population generation via combinatorial optimization

We built on Combinatorial Optimization code by Dr. Paul Williamson (link) to create synthetic cohorts for various populations in Boston. These synthetic cohorts have been used in a number of applications at BU: to assess the relationship between smoking and demographic variables in New Bedford, MA (Levy et. al., 2014), to assess impacts of multiple health promoting behaviors (diet, exercise) in combination with environmental exposures (Basra et. al., 2017), and testing the impact of neighborhood greening on birthweight (Milando et. al., 2021). A summary of the methods and current applications are described here.


miniCONTAM rapid indoor air health impact estimation

In progress

CONTAM Energy-plus automated coupled model preparation

In progress


Wald test

Wald test percentiles

Johnson and Romer 2016 provided an algorithm to compare a limited set of percentiles between two distributions. I created an R function to implement it.


DLNM sweep

DLNM parameter sweep

These functions allow you to test many crossbasis options in the DLNM framework.