This repository contains R-code for performing large-scale PCA eigenvector analysis using the singular-value decomposition as implemented by the RSpectra package, using TraCE-21ka data as an example. In addition, rolling correlation and PCA analysis using the roll package and rollapply() function in the zoo package is illustratred using a set of area-average time series for the mid-continent of North America. THe correlations and compoent loadings are plotted using the 'corrplot' and 'qgraph' packages.
The contents of the various folders are as follows
| Folder | Description |
|---|---|
/R |
R code |
/animations |
scripts for ImageMagick animated .gifs of the rolling correlation qgraphs |
/data |
data sets |
/figs |
figures |
/ncl |
NCL code for plotting component-loading maps and scores time series |