Add variable labels using the summarytools freq package

Question

I want to add variable labels to the frequency tables that I generated. But, I can't find that functionality in the summarytools documentation.

Here is my code:

the data

library(magrittr)
library(dplyr)
library(gtsummary)
library(summarytools)
require(pander)
library(knitr)
library(stringr)

data_in_na <- readr::read_table2('q1    q2
No  somelongresponsethattakesupmorethanonline--somelongresponsethattakesupmorethanonline
No  somelongresponsethattakesupmorethanonline--somelongresponsethattakesupmorethanonline
No  somelongresponsethattakesupmorethanonline--somelongresponsethattakesupmorethanonline
No  somelongresponsethattakesupmorethanonline--somelongresponsethattakesupmorethanonline
No  Yes
No  Always
Yes No
Yes No
Yes No
NA  NA
NA  NA
NA  NA
NA  NA
')


vct <- data_in_na %>% names(.)

function to create frequencies

create_freq <- function(v1) {
  #names <- c("var name 1", "var name 2")
  
  freq(v1,
       cumul = FALSE, 
       totals = TRUE,
       display.type = FALSE,
       variable.label = v1, 
       missing ='missing', 
       report.nas = TRUE)
}

loop to run through all the variables

for (i in vct) {
  tbl <- create_freq(data_in_na[i])   # build gtsummary table
  print(tbl)       # print table
  cat('\n\n----------------------------------------------------------\n\n')
  # d <- table(data_in[i])
  # print(kable(d))
  
}

what I want

Any suggestions please??

How are you going to use the table? I ask because when you view the table with `view(tbl)` the column with variable names is labeled `q2`. It doesn't look like the console output version has that option. However, I would guess that you have another purpose for the tables. What end state are you looking for? — Kat, Jan 16 '22 at 17:54
@Kat I am processing this in a rmarkdown file (writing to html) using a print statement inside a loop code chunk! — NewBee, Jan 18 '22 at 19:37

Kat · Accepted Answer · 2022-01-19T20:59:31.787

This adds a "Variable" label about the field names and removes the double hashtags. I also adjusted the standard max-width of 940 px to 1000px -- however, if the actual long name is a lot longer than your example data, this won't be much help. I've provided comments (// precedes inline comments in JS) so that you can see what each area of the Javascript is doing. You don't have to do anything special for the Javascript to work (it's built-in in R Markdown).

Using your code as it already is, add the following in as a chunk after your code.

If you try to run this chunk inline, it won't produce output. However, you'll see the output when you render.

For the questions added to the question labels, I've edited this to provide two different ways to do this. One uses R. The other uses JS.

I would suggest using R, because it's native. But you may have your own reasons for not doing so. This violates good naming practices, but it could be used for just these tables (if you wanted).

The R version is OPTION 1. This code goes between your data collection and storing the variable names in vct. (I added vct to this code, actually.)

```{r optQuestions, include = F,eval=F}
# not sure how you have the questions stored
# let's say you have the questions stored in a vector
questions = c("How often do you exercies?", "Do you like apples?")
qLabs = names(data_in_na)

names(data_in_na) <- paste(qLabs, questions)

vct <- data_in_na %>% names(.)
```

Now for the JS - 2 lines equate to doing what paste is doing in the R chunk. Using one or the other, you won't see a difference with this data specifically. Although, if there are a lot of questions, you may find the R method preferable.

```{r styler,results='asis',engine='js'}

// search for class and tags
elem = document.querySelector('div.hereForMod > pre > code');
// remove hashtags
elem.innerHTML = elem.innerHTML.replace(/#{2}/g, '');

// I missed what you wanted for the questions -- 
// this is option TWO: // add // if you want to ignore them (or delete)
elem.innerHTML = elem.innerHTML.replace(/q1/g, 'q1 How often do you exercise?')
elem.innerHTML = elem.innerHTML.replace(/q2/g, 'q2 Do you like apples?')

// change max width
newMax = "max-width:1000px;";  //currently 940
elem2 = document.querySelector('.main-container'); // find what to change
styler = elem2.getAttribute("style");          // get current inline styles
if(styler==undefined || styler==null){ 
  styler="";
}
elem2.setAttribute("style", styler+newMax);    // set new style settings 

```

If you add eval=FALSE to the chunk, you can see what it changes when you don't use it.

If you want a word other than Variables, you have to adjust that value next to Freq. (Currently, that value is 12.) You'll also want to make sure that you maintain the same number of spaces between that word and Freq.

Let me know if you have any questions.

Before edit:

After edit:

thanks for this! This is not quite what I want... I want to add a unique label for each of the questions... so for example q1: "how often to you exercise" and for q2: "do you like apples". The location of the label does not matter, as long as it is some where above the referencing table... — NewBee, Jan 19 '22 at 20:04
Sorry about that! I've added a way to do this part of your request; two different ways to do it. — Kat, Jan 19 '22 at 21:00

Add variable labels using the summarytools freq package

the data

function to create frequencies

loop to run through all the variables

what I want

1 Answers1