Author: Andrew Larkin
Affiliation: Oregon State University, College of Public Health and Human Sciences
Principal Investigator: Perry Hystad
Date last modified: September 23rd, 2018

Summary
This github repository contains the scripts used to create a global NO2 land use regression model. Land use estimates were derived using python and ArcGIS. Variable selection and model development were developed using R Studio.

Repository Structure
Files are divided into three folders, with each folder corresponding to a unique stage of model development.

• variable estimates - scripts in this folder were used to derive NO2 estiamtes at air monitor locations using parallel processing.
  
• upwind estimates - scripts in this folder were used to estimate length of road upwind from air monitor locations.
• statistical analysis - scripts in this folder were used to perform lasso variable selection, model evaluation, and sensitivity analysis.
  

External Links

• Publication - https://pubs.acs.org/doi/abs/10.1021/acs.est.7b01148
• Model estimates and underlying datasources - https://health.oregonstate.edu/labs/spatial-health/resources-equipment