README.md

In joshyam-k/messy: What the Package Does (One Line, Title Case)

messy

A package that attempts to remedy some of the difficulties of cleaning data composed of survey inputs. Often times when trying to lump or clean up a categorical variable that was a survey input you’ll come across things like “The United States”, “the united states”, or “THE UNITES STATES”. It can be time consuming and frustrating to try to identify each capitalization and punctuation variant and not all of us are stringr masters. All that Messy does is gives a nice way to automate the process of testing for a lot of capitalization and punctuation variants in a variable.

The developmental version of the package can be installed through github using:

devtools::install_github("joshyam-k/messy")

How to use the package

I envision the main function in the package, messy::messy_string to be used inside of dplyr::case_when as follows.

Suppose we have the following data frame:

library(tibble)
library(dplyr)
library(messy)

survey <- tibble(
  country = c('The United States', "the United states", "the United-kingdom", "The United Kingdom"),
  ages = 24:27
)

We’d like to clean up our categorical variable and in this example we could do it by hand, but in a more large scale case it could get very tedious.

messy_string takes a string as an input and outputs capitalization and punctuation variants of that string. The function is most useful when used on a string that contains varying punctuation choices.

messy_string("the united-kingdom")

##  [1] "the united kingdom" "the united Kingdom" "the United kingdom"
##  [4] "the United Kingdom" "The united kingdom" "The united Kingdom"
##  [7] "The United kingdom" "The United Kingdom" "the united KINGDOM"
## [10] "the UNITED kingdom" "the UNITED KINGDOM" "THE united kingdom"
## [13] "THE united KINGDOM" "THE UNITED kingdom" "THE UNITED KINGDOM"
## [16] "the united-kingdom" "the united-Kingdom" "the United-kingdom"
## [19] "the United-Kingdom" "The united-kingdom" "The united-Kingdom"
## [22] "The United-kingdom" "The United-Kingdom" "the united-KINGDOM"
## [25] "the UNITED-kingdom" "the UNITED-KINGDOM" "THE united-kingdom"
## [28] "THE united-KINGDOM" "THE UNITED-kingdom" "THE UNITED-KINGDOM"

As you can see, when inputting a string with some type of punctuation the output contains capitalization variants both with and without the punctuation! Thus when using the function to deal with punctuation variety you should always input the string with as much punctuation as possible.

The function can then be used directly inside of case_when for a much easier cleaning process!

survey %>% 
  mutate(
    country_clean = case_when(
      country %in% messy_string("the united states") ~ "USA",
      country %in% messy_string("united-kingdom") ~ "UK"
    )
  )

## # A tibble: 4 x 3
##   country             ages country_clean
##   <chr>              <int> <chr>        
## 1 The United States     24 USA          
## 2 the United states     25 USA          
## 3 the United-kingdom    26 <NA>         
## 4 The United Kingdom    27 <NA>

One final note is that every problem that you could solve with messy could also be done with stringr and regular expressions. I wrote this package fully aware of this fact, and acknowledge that there will be a lot of places where this package is far less usefull or elegant than a stringr approach. Nonetheless, messy provides a way to perhaps more intuitively try to deal with all the messiness of survey inputted data.

joshyam-k/messy documentation built on Dec. 21, 2021, 2:21 a.m.