Running Mann-Kendall on multiple columns in r

user_14296293 0
Viewed 301

I am new in R and would like to run a mann-kendall on multiple columns at once.

structure(list(Year = c(1997, 1999, 2001, 2002), pH = c(8, 8.4, 
  8.2375, 8.27333333333333), Colour = c(16, 50.5, 21, 17.9090909090909
  )), row.names = c(NA, -4L), class = c("tbl_df", "tbl", "data.frame"
))

This is a sample of my data

Here is what I tried for an individual column

MannKendall(NoordAnnual$Colour)
# tau = -0.137, 2-sided pvalue =0.4173

I am hoping the get a table with the values of tau and p-value for all columns.

r
1 Answers

We can use lapply to loop over the columns of interest. Here, the first column is dropped as it is 'Year'

library(Kendall)     
out <- lapply(NoordAnnual[-1], MannKendall)
out
#$pH
#tau = 0.333, 2-sided pvalue =0.7341

#$Colour
#tau = 0, 2-sided pvalue =1

Or with dplyr

library(dplyr)
NoordAnnual %>%
       summarise(across(-1,  ~list(MannKendall(.))))

If we want as a table

library(tidyr)
library(broom)
NoordAnnual %>%
      summarise(across(-1,  ~list(MannKendall(.) %>%
               tidy %>%
               select(p.value, statistic)))) %>%
      pivot_longer(everything()) %>%
      unnest(c(value))
# A tibble: 2 x 3
#  name   p.value statistic
#  <chr>    <dbl>     <dbl>
#1 pH       0.734     0.333
#2 Colour   1         0
user_37322710
Related