Using R in your text

library(tidyverse) # you know why ;-)
library(broom)   # for tidying model output
library(glue)    # for pasting character values
library(sjPlot)  # for creating tables of models

Loading penguins data

library(palmerpenguins)
data("penguins")

Some information about the data

The penguins data contains information of 344 penguins. The number of penguins is ‘calculated’ by R in the text. Hence, if you would go back to the islands and measure some additional penguins. You can just rerun the script and the count would adjust itself…

The penguins live on three islands. We want to know how many penguins were observed on each island. So, we apply some of our tidyverse skills.

## Count by group
islands <- penguins %>%
  group_by(island) %>%
  count() 
islands

# A tibble: 3 × 2
# Groups:   island [3]
  island        n
  <fct>     <int>
1 Biscoe      168
2 Dream       124
3 Torgersen    52

Now, we can refer to the values in this data set by indexing! For example, there live 168 penguins on Biscoe. This is easy in this case (because the data set is small), but becomes difficult with larger data sets and you can easily make a mistake.
To make youR life easier, the package printy can be used. (More information can be found at https://www.tjmahr.com/lists-knitr-secret-weapon/)

# install.packages("remotes")
# remotes::install_github("tjmahr/printy")
library(printy)

## Create a list
islands_list <- super_split(islands, island)

# Also possible
#islands <- penguins %>%
#  group_by(island) %>%
#  count() %>%
#  super_split(island)

Doing this, we converted the data set to a list of values for each island.

islands_list

$Biscoe
# A tibble: 1 × 2
# Groups:   island [1]
  island     n
  <fct>  <int>
1 Biscoe   168

$Dream
# A tibble: 1 × 2
# Groups:   island [1]
  island     n
  <fct>  <int>
1 Dream    124

$Torgersen
# A tibble: 1 × 2
# Groups:   island [1]
  island        n
  <fct>     <int>
1 Torgersen    52

Now, we can refer to the values by using $ and the name of the island. For example, the number of penguins on Torgersen can be found by using the following code: islands_list$Torgersen$n. Reporting on the number of penguins on each island now become easy.

The penguins live on three islands: Biscoe (n = 168), Dream (n = 124), and Torgersen (n = 52).

Fitting a linear model

We fit a linear model with body_mass_g as dependent variable and species and flipper_length_mm as independent variable

## Fit a linear model
Model_lm <- lm(body_mass_g ~ flipper_length_mm + species, data = penguins)

Here is a nice table with the estimates of the model we fitted.

tab_model(Model_lm)

	body mass g
Predictors	Estimates	CI	p
(Intercept)	-4031.48	-5180.51 – -2882.45	<0.001
flipper length mm	40.71	34.66 – 46.75	<0.001
species [Chinstrap]	-206.51	-320.07 – -92.95	<0.001
species [Gentoo]	266.81	79.43 – 454.19	0.005
Observations	342
R2 / R2 adjusted	0.783 / 0.781

If we discuss the results of this model we want to refer to the estimates and confidence intervals. To do so, we create a list that contains 4 elements (one for each estimate):
- Intercept
- flipper_length_mm
- speciesChinstrap
- speciesGentoo

list_lm <- tidy(Model_lm, conf.int = TRUE) %>%
  ## Remove the brackets around the Intercept-term
  mutate(term = ifelse(term == "(Intercept)", yes = "Intercept", no = term)) %>%
  ## Round the values to 2 decimals
  mutate_at(vars(estimate, conf.low, conf.high), fmt_fix_digits, 2) %>%
  ## Create a new variable 'ci' by glueing the lower and upper limit
  mutate(ci = glue("[{conf.low}", "{conf.high}]")) %>%
  select(term, estimate, ci) %>%
  ## Apply the super_split function
  super_split(term)

Now, we can refer to the estimate of the intercept by using list_lm$Intercept$estimate and to the confidence interval by using list_lm$Intercept$ci.

The intercept of the model is -4031.48 (95% CI = [-5180.51-2882.45])