Application of R software in data analysis of environmental epidemiology: health effects of air pollution

Objective To implement individual exposure assessment of air pollution based on personal address information using the R language tidyverse package and exchange experience in the use of the method.

Methods The data of cardiovascular and cerebrovascular mortality in Nanjing from 2017 to 2019 were simulated with computer, and the meteorological and environmental pollutant monitoring data in the same period were obtained online from the network. The data then were filtered, connected, and summarized through dplyr package in the R language tidyverse package, and then deformed and converted by the tidyr package, and achieved traversal loops by the purrr package. The nearest environmental monitoring sites exposure and inverse distance weighted exposure were calculated by latitude and longitude method.

Results Using the crawler technology of the rvest package meteorological data, environmental pollutant monitoring data and others were obtained, and using tidy and purrr packages for data cleaning, using geosphere packages to process spatial data, to assess the individual exposure by calculating the nearest site and inverse distance interpolation.

Conclusion Compared with the base package, the R language tidyverse has the advantages of consistent syntax, efficient data processing ability, and being easy to master. It could be improved effectively by using tidyverse for data cleaning, summary statistics, exposure calculation and other data processing in environmental epidemiological studies. This study provided the code for data processing by using the R language tidyverse package for inverse distance weighting calculation, and realized a method to evaluate individual daily air pollutants exposure, which provided an effective tool for conducting air pollutants exposure assessment.