This repository holds the functions and data visualisations developed for our ACP paper, implemented in R.
The functions cover a range of standard univariate and multivariate data analysis, with the data plotting as used for all paper figures.
Most of the functions available here can be used for similar datasets; some hard coding is present, and we will make the package fully generalisable as soon as we can.
Package dependencies can be found here, our credit and thanks to the authors of these!
Some of the functions include:
-
univariate statistics: Spearman correlations and p-values, in a handy text output
-
multivariate statistics: principal components analysis (currently without cross-validation), plus scores and loadings plots. These plots can also be used for PLSR models, which can be constructed using the pls package (also in R).
- correlation heatmaps per assay and per season
- and the MLR optimisation and model visualisations
If you find any of the code useful or have suggestions for improvement, please let us know!
Atmospheric conditions and composition that influence PM2.5 oxidative potential in Beijing, China
Campbell, Wolfer et al.
Atmos. Chem. Phys., 21, 5549–5573, 2021
https://doi.org/10.5194/acp-21-5549-2021
https://acp.copernicus.org/articles/21/5549/2021/acp-21-5549-2021.html