Error in diversity(x, index = "shannon") : input data must be numeric

[Anibaba007]

Hello everyone,
I have tried running the analysis which aims to plot Shannon diversity index across land-use types.
However, upon running the H line I got the error message:

H <- apply(nativeAB, MARGIN = 1,FUN = diversity, index="shannon")
Error in FUN(newX[, i], ...) : input data must be numeric

Below are my script and data attached.

nativeAB = read.csv('all_transposedDR.csv')#Read data
head(nativeAB)#look at first 6 rows
str(nativeAB)

A plot of Shannon-Wiener H' across the three land-use types

library(vegan)

H <- apply(nativeAB, MARGIN = 1,FUN = diversity, index="shannon")
H
CI<- apply(nativeAB, MARGIN = 1, FUN = H_CI)
CI
diversitybarplot <- barplot(shannon, ylim = c(0, max(H+CI)), cex.names = 0.8)
arrows(diversitybarplot, H+CI, diversitybarplot, H-CI, length = 0.1, angle = 90, code = 3)
title(xlab = "land use types", ylab = "Shannon-Wiener H' index")

Copy of all_transposedDR.xlsx

[jarioksa]

I think the error message is clear: input data must be numeric. When I look at the Excel file that you sent, I see that input data are not numeric, and you should have seen the same when you issued head(nativeAB). At least variables site and land.use are character strings (non-numeric) and also variables transect and plot should be analysed as non-numeric factors. You should omit those columns from your calculation. This is also very important: probably transect and plot will be read in as numeric and diversity can use them, but if you include them in your diversity calculations, your results are rubbish.

You can remove columns using negative indexing (and you do not need to use apply: diversity is automatically applied to each row):

H <- diversity(nativeAB[, -(1:4)]) # removes cols 1 to 4, but check the actual indexing in your data

As a non-vegan issue: site and land.use are better treated as factors than character strings. However, for some obscure reason, R version 4 handles them as character strings. If you are using current R, you should add keyword stringsAsFactors=TRUE in your read.csv() command. Then you should still change your numeric transect and plot to factors.

The last rows of your example are a mystery to me, as I have no idea what is function H_Cl or how you get something that you call shannon in your barplot() command.