maybe not up to date.
not used for analysis.
note:
This file requires the functions dataset raw_functions
which is constructed in the previous scripts 1read_raw_dataset.Rmd
and 2calc_raw_dataset.Rmd
. Here, it is named raw_grlfuns
.
library(naniar)
A selection of the functions which are included in the BetaDivMultifun analysis is stored in grlfuns
.
vis_miss(grlfuns[, !colnames(grlfuns) %in% c("Explo","Plotn", "Plot"), with=F]) # plot and explo don't contain any missing data
Values with over 21% missing data are excluded.
#USER : set treshold for missing values treshold <- 0.21 t <- apply(grlfuns, 2, function(x) sum(is.na(x))) exclude <- names(which(t > 150 * treshold)) small_grlfuns <- grlfuns[, !colnames(grlfuns) %in% exclude, with=F]
Visualisation of new situation
vis_miss(small_grlfuns[, !colnames(grlfuns) %in% c("Explo","Plotn", "Plot"), with=F])
Are there plots which are missing in all variables?
gg_miss_upset(small_grlfuns)
No plots are missing in all variables.
M <- cor(small_grlfuns[, !colnames(grlfuns) %in% c("Explo","Plotn", "Plot"), with=F], use="pairwise.complete.obs") corrplot::corrplot(M,type="lower",addCoef.col = "black",method="color",diag=F, tl.srt=1, tl.col="black", mar=c(1,0,1,0), number.cex=0.6)
Correlations for analysis should not be over 0.7.
treshold <- 0.7 M[M < 0.7 & M > -0.7] <- 0 corrplot::corrplot(M,type="lower", tl.srt=1, tl.col="black", diag = F, title = "Correlations over 0.7", mar=c(0,0,2,0))
High correlations between following variables :
Either Soil.C.stock or SoilOrganicC.
vis_miss(small_grlfuns[, colnames(small_grlfuns) %in% c("Soil.C.stock","SoilOrganicC"), with=F])
Soil Organic C is probably contained in Soil C stock, as the stock consists of both organic and anorganic C. Are both available to the ecosystem? Who consumes it? Microbes? Plants? animals?
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.