STAT 118: Notes D

Pretty tables with kableExtra

Author
Affiliation

Emily Malcolm-White

Middlebury College 

#LOAD PACKAGES 
library(tidyverse)
#LOAD DATA 
library(palmerpenguins)
data(penguins)

#CLEAN UP DATA
penguins <- penguins %>%
  drop_na()
#sometimes this is appropriate. It's questionable here...

mutate

The mutate function allows you create a new column which is a function of other columns. This can be useful to converting units.

Artwork by @allisonhorst

For example, let’s calculate create a new column which displays the body length weight in pounds (lbs) instead of grams. Recall: to convert from grams to pounds we need to multiply by 0.00220462

penguins <- penguins %>% 
  mutate(body_mass_lbs = body_mass_g*0.00220462)

This can also be useful for making new calculations based on existing data (for example, price and number of square feet could be used to calculate price per square foot).

case_when

Case when can be used in combination with mutate to create a new column with a categorical variable conditional on the values in another column.

Artwork by @allisonhorst

For example:

penguins <- penguins %>% 
  mutate(penguin_length_cat = case_when(bill_length_mm > 50 ~ 'whoa! huge bill!', TRUE ~ '--' ))
Tip

For those of you who have taken a computer science class before, you may notice that case_when is similar to using an ifelse statement. You can also use ifelse in R if you’d prefer!

penguins <- penguins %>% 
  mutate(penguin_length_cat = ifelse(bill_length_mm > 50, 'whoa! huge bill!', '--' ))

default printing style of a table

Let’s consider our table from last class:

penguins %>%
  group_by(species) %>%
  summarise(average_bill_lenth = mean(bill_length_mm), 
            average_bill_depth = mean(bill_depth_mm))
# A tibble: 3 × 3
  species   average_bill_lenth average_bill_depth
  <fct>                  <dbl>              <dbl>
1 Adelie                  38.8               18.3
2 Chinstrap               48.8               18.4
3 Gentoo                  47.6               15.0

When we knit this up it looks like of ugly…

Using kable to get pretty tables

library(kableExtra)
penguins %>%
  group_by(species) %>%
  summarise(average_bill_length = mean(bill_length_mm), 
            average_bill_depth = mean(bill_depth_mm)) %>% 
  kbl()
species average_bill_length average_bill_depth
Adelie 38.82397 18.34726
Chinstrap 48.83382 18.42059
Gentoo 47.56807 14.99664 
#OR 

table1 <- penguins %>%
  group_by(species) %>%
  summarise(average_bill_length = mean(bill_length_mm), 
            average_bill_depth = mean(bill_depth_mm))

kbl(table1)
species average_bill_length average_bill_depth
Adelie 38.82397 18.34726
Chinstrap 48.83382 18.42059
Gentoo 47.56807 14.99664

Options in kable

We customize the content so it’s displaying the information more clearly.

table1 %>%
  kbl(col.names = c("Species", "Average Bill Length", "Average Bill Depth"), 
    caption = "Average Bill Characteristics by Species", 
    digits = 2)
Average Bill Characteristics by Species
Species Average Bill Length Average Bill Depth
Adelie 38.82 18.35
Chinstrap 48.83 18.42
Gentoo 47.57 15.00

Better…

pretty styling

table1 %>%
  kbl(col.names = c("Species", "Average Bill Length", "Average Bill Depth"), 
    caption = "Average Bill Characteristics by Species", 
    digits = 2) %>%
  kable_styling()
Average Bill Characteristics by Species
Species Average Bill Length Average Bill Depth
Adelie 38.82 18.35
Chinstrap 48.83 18.42
Gentoo 47.57 15.00

Many options for customizing the look of the tables – more here: https://cran.r-project.org/web/packages/kableExtra/vignettes/awesome_table_in_html.html

Let’s try one…

Let’s make each row’s color correspond to

table1 %>%
  kbl(col.names = c("Species", "Average Bill Length", "Average Bill Depth"), 
    caption = "Average Bill Characteristics by Species", 
    digits = 2) %>%
  kable_paper() %>%
  column_spec(1, bold=T) %>%
  row_spec(2, color = "#c85bcc") %>%
  row_spec(3, color = "#067176") %>%
  row_spec(1, color = "#ff7501")
Average Bill Characteristics by Species
Species Average Bill Length Average Bill Depth
Adelie 38.82 18.35
Chinstrap 48.83 18.42
Gentoo 47.57 15.00

RStudio hosts a table contest every year!

Tip

As a general rule, you should have the content of the table as you’d like it (the exact columns and rows you want) first and then you can make it pretty using the kableExtra package.