format_values: Format values for table input

In LDSamson/associationstudies: Custom functions useful to perform association studies and prediction analyses

Format values for table input

Description

Formats the results of a summary value so that it can be input into a (latex) table, such as the output of knitr::kable(). Aims to give a publication-ready result. Expected input is a numeric vector, with the summary statistic value on the first place, and on the second/third place the standard deviation or confidence interval. It is designed to work with the output of functions such as gm_mean, median_iqr and mean_sd.

Usage

format_values(x, n.digits = 3, latex_output = TRUE, na.rm = FALSE, ...)

Arguments

x	numeric values that need to be summarized (for now, only geometric mean + confidence interval is supported)
n.digits	digits to preserve in output
latex_output	Logical, Whether dollar signs should be used around the output string, for latex formatting in knitr::kable. Make sure to set escape=FALSE in knitr::kable function when using this.
na.rm	logical, indicating whether NA values should be excluded
...	other parameters, parsed onto function gm_mean

Details

Inspired by this post.

Value

Character value to input in kable function

Examples

df <- data.frame(
  V1 = abs(rnorm(n = 1000, mean = 10000, sd = 10000)),
  V2 = exp(rnorm(n = 1000, mean = log(10^6), sd = log(10^4))),
  V3 = abs(rnorm(n = 1000, mean = 100, sd = 10))
)

#V1: median (IQR)
format_values(median_iqr(df$V1), n.digits = 1)
#V2: geomean (conf)
format_values(gm_mean(df$V2), n.digits = 1)
#V3: mean (SD)
format_values(mean_sd(df$V3), n.digits = 1)

## illustrate how to use summary tables:
library(dplyr)
library(knitr)
library(kableExtra)

# create dummy data, with some variables having much higher average values
# than others, to illustrate the use of this function:
table_data <- immune_data %>%
  mutate(across(c(Neutrophils, Monocytes), .fns = ~(.+1)*10^5))

summ_table <- immune_data %>%
  group_by(Sex) %>%
  summarize(
    Neutrophils = format_values(gm_mean(Neutrophils)),
    Tregs = format_values(gm_mean(Tregs))
  )
knitr::kable(summ_table) %>%
  kableExtra::kable_styling()

# The function is also helpful to create many summary values at once:
library(tidyr)
table_data_long <- table_data %>%
  select(-c(Sex, Batch, Frailty.index)) %>%
  tidyr::pivot_longer(col = everything())

summ_table_long <- table_data_long %>%
  group_by(name) %>%
  summarize(
    "Summary value" = format_values(gm_mean(value, offset = 1), n.digits = 2)
  )
# offset gives better results when geomean and confidence interval are calculated
# and there are many zeros in the data. See details of the function gm_mean()
# for more information.

knitr::kable(summ_table_long) %>%
  kableExtra::kable_styling()

LDSamson/associationstudies documentation built on April 14, 2025, 11:52 a.m.