4. Wrangling

library(PAGES)

Tidyverse packages

In this last exercise I show some functionality of the tidyr (Wickham 2021) package.

From the website:

"The goal of tidyr is to help you create tidy data. Tidy data is data where:

  • Every column is variable.
  • Every row is an observation.
  • Every cell is a single value.

Tidy data describes a standard way of storing data that is used wherever possible throughout the tidyverse."

This looks like this:

obs = observation, var = variable and val = value

obs = observation, var = variable and val = value

Kuhjoch wide format

The following format of the data frame (or tibble (Müller and Wickham 2021) in the tidyverse universe) is a format that we are used to see in publications. Here count data of all different elements each occupy their own column (i.e., variable).

#> # A tibble: 6 x 5
#>   height landplant_pollen landplant_spores environment_aquat… environment_terre…
#>    <dbl>            <int>            <int>              <int>              <int>
#> 1   1600               31              277                 65                308
#> 2   1365               27              262                114                289
#> 3   1125               22              282                 40                304
#> 4   1045               24              276                 44                300
#> 5    965               29              276                 75                305
#> 6    885               19              284                 70                303

However, we can also describe this data in a long format, as the class of the values in the cells for the palynomorphs are all the same (i.e., they are all integer count data). So, how can we make a long format data frame out of this?

Kuhjoch long format

For such operations, we can use the tidyr (Wickham 2021) function pivot_longer().

However, the question is a bit of trick question, as the variable names also contains information. Actually, there are additional variables stashed in the column names. In this case, the word in the variable name before the subscript _ is another categorical variable dividing the the label in two subsets; "landplant"; denoting whether the variable label describes a characteristic of land plant community composition (either pollen or spores), and "environment"; denoting whether the variable label describes a characteristic of the environment (i.e., is the palynomorph an aquatic or terrestrial component?).

To achieve the previous describe transformation to a long format data frame, and the extraction of the information in the column names, we construct the following call.

# long format
pivot_longer(
  data = kuhjoch, 
  cols = -height,  
  names_to = c("type", "label"),
  values_to = "count",
  names_sep = "_"
  )
#> # A tibble: 180 x 4
#>    height type        label       count
#>     <dbl> <chr>       <chr>       <int>
#>  1   1600 landplant   pollen         31
#>  2   1600 landplant   spores        277
#>  3   1600 environment aquatic        65
#>  4   1600 environment terrestrial   308
#>  5   1365 landplant   pollen         27
#>  6   1365 landplant   spores        262
#>  7   1365 environment aquatic       114
#>  8   1365 environment terrestrial   289
#>  9   1125 landplant   pollen         22
#> 10   1125 landplant   spores        282
#> # … with 170 more rows

Note that in this call we select column with -Height, so we take everything except the height variable. Also note that the vector of two elements for the names_to argument defines the two new categorical columns and name_sep controls how the original column names are split ("_").

References

Müller, Kirill, and Hadley Wickham. 2021. Tibble: Simple Data Frames. https://CRAN.R-project.org/package=tibble.
Wickham, Hadley. 2021. Tidyr: Tidy Messy Data. https://CRAN.R-project.org/package=tidyr.