‘[R] tables with {gt} and {gtExtras}’

Visuals

2022

Author

Cliff Weaver

Published

July 24, 2017

The {gt} extension already allowed to easily create tables from raw dataset, but now the {gtExtras} extension adds many customization options. Here we will illustrate the possibilities of these packages with TidyTuesday dataset on Tour de France riders, extracted from Alastair Rushworth’s {tdf} extension.

1. Tables with text

We will start by loading the data.

library(tidyverse)
# Load data: 
# Whole race winners
tdf_winners <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2020/2020-04-07/tdf_winners.csv')
# Stages winners
tdf_stages <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2020/2020-04-07/tdf_stages.csv')

head(tdf_stages)

# A tibble: 6 × 8
  Stage Date       Distance Origin            Destination   Type  Winner Winne…¹
  <chr> <date>        <dbl> <chr>             <chr>         <chr> <chr>  <chr>  
1 1     2017-07-01      14  Düsseldorf        Düsseldorf    Indi… Gerai… GBR    
2 2     2017-07-02     204. Düsseldorf        Liège         Flat… Marce… GER    
3 3     2017-07-03     212. Verviers          Longwy        Medi… Peter… SVK    
4 4     2017-07-04     208. Mondorf-les-Bains Vittel        Flat… Arnau… FRA    
5 5     2017-07-05     160. Vittel            La Planche d… Medi… Fabio… ITA    
6 6     2017-07-06     216  Vesoul            Troyes        Flat… Marce… GER    
# … with abbreviated variable name ¹Winner_Country

As a first example, we will make a table with the stage winners of a given year: 1961. We will see later why I chose this year…

tdf_61 <- tdf_stages%>%
  # Create year column from Date
  mutate(year = lubridate::year(Date))%>%
  # Keep only 1961 data
  filter(year==1961)%>%
  # Remove year column
  select(-year)

We will now load the {gt} package to format our first table. One of the main advantages of {gt} is that it fits perfectly into tidyverse pipes: with just have to add the gt() function to create a table.

# Load {gt}
library(gt)

# Make table with gt()
tab<-tdf_61 %>%
  # Keep only first 6 rows as example
  head(6)%>%
  # Make table
  gt() 

tab

Stage	Date	Distance	Origin	Destination	Type	Winner	Winner_Country
1a	1961-06-25	136.5	Rouen	Versailles	Plain stage	André Darrigade	FRA
1b	1961-06-25	28.5	Versailles	Versailles	Individual time trial	Jacques Anquetil	FRA
2	1961-06-26	230.5	Pontoise	Roubaix	Plain stage	André Darrigade	FRA
3	1961-06-27	197.5	Roubaix	Charleroi	Plain stage	Emile Daems	BEL
4	1961-06-28	237.5	Charleroi	Metz	Plain stage	Anatole Novak	FRA
5	1961-06-29	221.0	Metz	Strasbourg	Stage with mountain(s)	Louis Bergaud	FRA

Once the table is created, we can use the pipe operator %>% to add few functions to customize the table. Here are some examples of how to add a title or format the columns:

tab<-tab %>%
  # Add title and subtitle
  tab_header(
    title = "Stage winners",
    # Use markdown syntax with md()
    subtitle = md("Tour de France **1961**")
  )%>%
  # Fomat date without year information
  fmt_date(
    columns = Date,
    date_style = 9
  )%>%
  # Format distance without decimal 
  fmt_number(
    columns = Distance,
    decimals = 0,
    # Add 'km' as suffix
    pattern = "{x} km"
  )%>%
  # Rename column
  cols_label(
    Winner_Country = "Nationality"
  )

tab

Stage	Date	Distance	Origin	Destination	Type	Winner	Nationality
Stage winners
Tour de France 1961
1a	25 June	136 km	Rouen	Versailles	Plain stage	André Darrigade	FRA
1b	25 June	28 km	Versailles	Versailles	Individual time trial	Jacques Anquetil	FRA
2	26 June	230 km	Pontoise	Roubaix	Plain stage	André Darrigade	FRA
3	27 June	198 km	Roubaix	Charleroi	Plain stage	Emile Daems	BEL
4	28 June	238 km	Charleroi	Metz	Plain stage	Anatole Novak	FRA
5	29 June	221 km	Metz	Strasbourg	Stage with mountain(s)	Louis Bergaud	FRA

If you like customization, the {gt} extension allows to change the appearance of each cell (color, background, font, border…). Using the tab_style() function, you may customize the title as follows:

tab %>%
  tab_style(
    # Select object to modify
    locations = cells_title(groups = 'title'),
    # Specify text style
    style = list(
      cell_text(
        font=google_font(name = 'Bebas Neue'), 
        size='xx-large',
        color='indianred'
  )))

Stage	Date	Distance	Origin	Destination	Type	Winner	Nationality
Stage winners
Tour de France 1961
1a	25 June	136 km	Rouen	Versailles	Plain stage	André Darrigade	FRA
1b	25 June	28 km	Versailles	Versailles	Individual time trial	Jacques Anquetil	FRA
2	26 June	230 km	Pontoise	Roubaix	Plain stage	André Darrigade	FRA
3	27 June	198 km	Roubaix	Charleroi	Plain stage	Emile Daems	BEL
4	28 June	238 km	Charleroi	Metz	Plain stage	Anatole Novak	FRA
5	29 June	221 km	Metz	Strasbourg	Stage with mountain(s)	Louis Bergaud	FRA

You may now try to customize a few parts of the table according to your taste! To help you select the elements you want to modify, here is a table summarizing the different parts of the table:

# out.width = "300px", echo = FALSE}
knitr::include_graphics("parts.png")

You need to specify the name of the element of the table that you want to modify, after “cells_”, in the locations argument. These helper functions to target the location cells associated with the styling are summarized here.

But, as you can see, it can be tedious to modify each part of the table in this way. Fortunately, the new {gtExtras} extension allows to use predefined themes.

# Install gtExtras
# remotes::install_github("jthomasmock/gtExtras")

# Load extension
library(gtExtras)
# Apply 'New York Times' theme
tab<-tab%>%
  gtExtras::gt_theme_nytimes()

tab

Stage	Date	Distance	Origin	Destination	Type	Winner	Nationality
Stage winners
Tour de France 1961
1a	25 June	136 km	Rouen	Versailles	Plain stage	André Darrigade	FRA
1b	25 June	28 km	Versailles	Versailles	Individual time trial	Jacques Anquetil	FRA
2	26 June	230 km	Pontoise	Roubaix	Plain stage	André Darrigade	FRA
3	27 June	198 km	Roubaix	Charleroi	Plain stage	Emile Daems	BEL
4	28 June	238 km	Charleroi	Metz	Plain stage	Anatole Novak	FRA
5	29 June	221 km	Metz	Strasbourg	Stage with mountain(s)	Louis Bergaud	FRA

Seven themes are available with {gtExtras}.

We are almost done with this first table. Now it is time to see why I select the stage winnrs of year 1961 as an example. With gt_highlight_rows(), we may highlight the name of one rider.

tab%>%
  gtExtras::gt_highlight_rows(
    # Row to highlight
    rows = 5, 
    # Background color
    fill = "lightgrey",
    # Bold for target column only
    bold_target_only = TRUE,
    # Select target column
    target_col = Winner
  )

Stage	Date	Distance	Origin	Destination	Type	Winner	Nationality
Stage winners
Tour de France 1961
1a	25 June	136 km	Rouen	Versailles	Plain stage	André Darrigade	FRA
1b	25 June	28 km	Versailles	Versailles	Individual time trial	Jacques Anquetil	FRA
2	26 June	230 km	Pontoise	Roubaix	Plain stage	André Darrigade	FRA
3	27 June	198 km	Roubaix	Charleroi	Plain stage	Emile Daems	BEL
4	28 June	238 km	Charleroi	Metz	Plain stage	Anatole Novak	FRA
5	29 June	221 km	Metz	Strasbourg	Stage with mountain(s)	Louis Bergaud	FRA

Anatole Novak is my grand father. He won one Tour de France stage in 1961, so this is the reason I chose this year as an example. In his career, he also helped Anquetil win several Tour de France. We will see how many in the next table!

2. Add images to table

In this second part, we will see more formatting options, in particular how we can add images to a table.

We will now look at the riders who have won the most Tour de France. This information may be extracted from the tdf_winners dataset:

# Data preparation:
most_wins<-tdf_winners%>%
  # Remove Armstrong (convicted for drug use)
  filter(winner_name!="Lance Armstrong")%>%
  # Keep only one spelling for Indurain
  mutate(winner_name=case_when(
    winner_name=='Miguel Induráin'~'Miguel Indurain',
    TRUE~winner_name
  ))%>%
  # Add variable to count titles
  mutate(ct=1)%>%
  # Group by winner name
  group_by(winner_name)%>%
  summarize(
    # Count titles
    Titles=sum(ct),
    # Add nationality
    Country=nationality[1],
    # Add nickname
    Nickname=nickname[1])%>%
  # Keep only winners with 3 titles or more
  filter(Titles>2)%>%
  # Sort by descending order
  arrange(-Titles)

most_wins

# A tibble: 8 × 4
  winner_name      Titles Country        Nickname                               
  <chr>             <dbl> <chr>          <chr>                                  
1 Bernard Hinault       5  France        Le Blaireau (The Badger), Le Patron (T…
2 Eddy Merckx           5  Belgium       The Cannibal                           
3 Jacques Anquetil      5  France        Monsieur Chrono, Maître Jacques        
4 Miguel Indurain       5  Spain         Miguelón, Big Mig (English)            
5 Chris Froome          4  Great Britain Froomey                                
6 Greg LeMond           3  United States L’Americain, LeMonster                 
7 Louison Bobet         3  France        Louison, Zonzon                        
8 Philippe Thys         3  Belgium       Le basset (The Basset Hound)

We are almost ready to convert this dataset into a {gt} table. Before that, we will rename the first column and reorder the others. Thus, the operations carried out on the data tibble will be reproduced on the {gt} table, which will keep the same column order. We will also clean up the list of nicknames.

# Data preparation:
most_wins<-most_wins%>%
  # Ordering columns
  select(
    Rider=winner_name,
    Nickname,Country,Titles)%>%
  # Cleaning nicknames
  mutate(Nickname=case_when(
    str_detect(Rider,'Hinault')~'The Badger',
    str_detect(Rider,'Anquetil')~'Maître Jacques',
    str_detect(Rider,'Indurain')~'Miguelón',
    str_detect(Rider,'LeMond')~"The American",
    str_detect(Rider,'Bobet')~'Zonzon',
    str_detect(Rider,'Thys')~'The Basset Hound',
    TRUE~Nickname
  ))

Preparation of the dataset is now complete, we may create the table. At the same time, we will discover a new function, gt_merge_stacks(), which allows to merge two columns (here riders name and nickname).

most_wins%>%
  gt()%>%
  tab_header(
    title = "Most sucessful riders in the Tour de France"
  )%>%
  gtExtras::gt_theme_nytimes()%>%
  # Merge riders' name and nickname on same column
  gtExtras::gt_merge_stack(col1 = Rider, col2 = Nickname)

Rider	Country	Titles
Most sucessful riders in the Tour de France
Bernard Hinault The Badger	France	5
Eddy Merckx The Cannibal	Belgium	5
Jacques Anquetil Maître Jacques	France	5
Miguel Indurain Miguelón	Spain	5
Chris Froome Froomey	Great Britain	4
Greg LeMond The American	United States	3
Louison Bobet Zonzon	France	3
Philippe Thys The Basset Hound	Belgium	3

The next column to customize is now the riders nationality: we will replace the name of the countries by an image of their respective flags. We can not do this directly from the table, we have to go back to the dataset to replace the name of the countries by the link to the flag image.

most_wins <- most_wins%>%
  mutate(Country = case_when(
    str_detect(Country,'France') ~ 'https://raw.githubusercontent.com/BjnNowak/TdF/main/fr.png',
    str_detect(Country,'Belgium') ~ 'https://raw.githubusercontent.com/BjnNowak/TdF/main/be.png',
    str_detect(Country,'Great Britain') ~ 'https://raw.githubusercontent.com/BjnNowak/TdF/main/uk.png',
    str_detect(Country,'Spain') ~ 'https://raw.githubusercontent.com/BjnNowak/TdF/main/sp.png',
    str_detect(Country,'United States') ~ 'https://raw.githubusercontent.com/BjnNowak/TdF/main/us.png'
  ))

Flag images can then be displayed in the table with the gt_img_rows() function.

tab2 <- most_wins%>%
  gt()%>%
  tab_header(
    title = "Most sucessful riders in the Tour de France"
  )%>%
  gtExtras::gt_theme_nytimes()%>%
  gtExtras::gt_merge_stack(col1 = Rider, col2 = Nickname)%>%
  # Add flag images
  gtExtras::gt_img_rows(columns = Country, height = 20)

tab2

Rider	Country	Titles
Most sucessful riders in the Tour de France
Bernard Hinault The Badger		5
Eddy Merckx The Cannibal		5
Jacques Anquetil Maître Jacques		5
Miguel Indurain Miguelón		5
Chris Froome Froomey		4
Greg LeMond The American		3
Louison Bobet Zonzon		3
Philippe Thys The Basset Hound		3

The last column to customize is the one with the number of titles. We could replace it with a barplot with gt_plt_bar(), but we will see the use of graphs in the third and last part. Here, we will replace the number of titles by icons representing a yellow jersey, which designates the leader of the Tour de France. To do so, we will use the gt_fa_repeats() function.

tab2%>%
  gtExtras::gt_fa_repeats(
    column=Titles,
    palette = "orange",
    name = "tshirt",
    align='left'
  )

Rider	Country	Titles
Most sucessful riders in the Tour de France
Bernard Hinault The Badger
Eddy Merckx The Cannibal
Jacques Anquetil Maître Jacques
Miguel Indurain Miguelón
Chris Froome Froomey
Greg LeMond The American
Louison Bobet Zonzon
Philippe Thys The Basset Hound

3. Add plots to table

To start this third part of our tutorial, we will create a new column showing a ‘title timeline’ for each rider. To make a plot from a column, it is generally necessary that this column contains several information. It is therefore necessary to group these data in a list, which will then be called by the function that will create the graph. Moreover, we want these lists to be ‘complete’, so that the data is homogeneous between the different runners. In this case, we will create a variable a variable that takes the value 1 for years with a title, 0 otherwise. To do so, we will use the complete() function. Finally, we will group the data in a list.

# Create a vector with names of riders with most wins 
names_most_wins<- most_wins %>%
  pull(Rider)

year_wins<-tdf_winners%>%
  # Rider column with one spelling for Indurain
  mutate(Rider=case_when(
    winner_name=='Miguel Induráin'~'Miguel Indurain',
    TRUE~winner_name
  ))%>%
  # Add ct variable to count years, 
  # with 1 for year with a title 
  mutate(ct=1)%>%
  # ... and create new rows with ct=0 
  # for years with no title 
  complete(Rider, edition, fill = list(ct = 0))%>%
  group_by(Rider)%>%
  # Create list for each rider
  summarise(Timeline = list(ct))%>%
  filter(Rider %in% names_most_wins)

year_wins

# A tibble: 8 × 2
  Rider            Timeline   
  <chr>            <list>     
1 Bernard Hinault  <dbl [106]>
2 Chris Froome     <dbl [106]>
3 Eddy Merckx      <dbl [106]>
4 Greg LeMond      <dbl [106]>
5 Jacques Anquetil <dbl [106]>
6 Louison Bobet    <dbl [106]>
7 Miguel Indurain  <dbl [106]>
8 Philippe Thys    <dbl [106]>

Once the list is created, we may plot the timeline in a table with gt_sparkline(). With the latest versions of the package, labels may be displayed on the x-axis but they may be turned off with label=FALSE.

# remotes::install_github("jthomasmock/gtExtras")
year_wins%>%
  gt()%>%
  gtExtras::gt_plt_sparkline(
    # Select column with data
    Timeline,
    # Color for min/max points
    #range_colors=c("#ABB4C4","#ef233c"),
    #palette = c("#ABB4C4","#ef233c"),
    palette = c("black", "black", "purple", "green", "lightgrey")
    # Line color
    #line_color="#DBDFE6",
    # Hide labels 
    # (for latest versions of {gtExtras}) only:
    # label=FALSE
    )%>%
  tab_header(title = "Titles timeline")%>%
  gtExtras::gt_theme_nytimes()

Rider	Timeline
Titles timeline
Bernard Hinault
Chris Froome
Eddy Merckx
Greg LeMond
Jacques Anquetil
Louison Bobet
Miguel Indurain
Philippe Thys

As there have been more than 100 editions of the Tour de France, it is difficult to read the timeline precisely, but we can at least compare the periods of activity each rider. For example, Thys was the first to win at least three tours, while Froome is the one who reached this threshold most recently.

As a last example, we will create a barplot with the number of stages won by each rider, as well as the type of stage won (Mountain stage, plain stage or time trial).

most_stages<- tdf_stages %>%
  mutate(Rider=case_when(
    Winner=='Miguel Induráin'~'Miguel Indurain',
    TRUE~Winner
  ))%>%
  filter(Rider %in% names_most_wins)%>%
  # Keep only 3 types of stages:
  # Time trial, mountain or plain
  mutate(TypeClean = case_when(
    str_detect(Type,"trial")~"Time trial",
    str_detect(Type,"mountain")~"Mountain stage",
    str_detect(Type,"Mountain")~"Mountain stage",
    str_detect(Type,"Hilly")~"Mountain stage",
    TRUE~"Plain stage"
  ))%>%
  group_by(Rider,TypeClean) %>%
  mutate(ct=1) %>%
  summarize(
    Wins=sum(ct)
  )%>%
  ungroup()%>%
  # Complete with NA for empty couples {rider*type of stages} 
  complete(Rider, TypeClean, fill = list(Wins = NA)) %>% 
  group_by(Rider)%>%
  summarise(Stages = list(Wins))

Again, once the list is created, we may plot the number of stages won in a table with gt_plt_bar_stack().

# Set color palette
pal_stages <- c('#264653','#e9c46a','#e76f51')

most_stages %>%
  gt()%>%
  gt_plt_bar_stack(
    # Column with data
    column=Stages,
    # Stacked barplot 
    position = 'stack', 
    # Set labels and color
    labels = c("Mountain stage", "Plain stage", "Time trial"),
    palette = pal_stages,
    # Barplot width
    width = 60#,
    # Same size for all labels
    #trim=TRUE
  )%>%
  tab_header(title = "Stages won")%>%
  gt_theme_nytimes()

Rider	Mountain stage\|\|Plain stage\|\|Time trial
Stages won
Bernard Hinault
Chris Froome
Eddy Merckx
Greg LeMond
Jacques Anquetil
Louison Bobet
Miguel Indurain
Philippe Thys

Finally, to group all this information in the same table, we will proceed with joins before creating the table, and then simply use the same functions as previously done to format the columns. It is also a good way to summarize what we have learned.

tab3<-most_wins%>%
  # Join tables
  left_join(year_wins)%>%
  left_join(most_stages)%>%
  # Make table
  gt()%>%
  # Set title
  tab_header(
    title = "Most sucessful riders in the Tour de France"
  )%>%
  # Set theme
  gtExtras::gt_theme_nytimes()%>%
  # Merge riders' name and nickname on same column
  gtExtras::gt_merge_stack(col1 = Rider, col2 = Nickname)%>%
  # Add flag images
  gtExtras::gt_img_rows(columns = Country, height = 20)%>%
  # Add yellow jerseys
  gtExtras::gt_fa_repeats(
    column=Titles,palette = "orange",
    name = "tshirt",align='left'
  )%>%
  # Format timeline
  gtExtras::gt_plt_sparkline(Timeline, 
                             #range_colors=c("#ABB4C4","#ef233c")
                             palette = c("black", "black", "purple", "green", "lightgrey")
                             #line_color="#DBDFE6"
  )%>%
  # Format stages won
  gt_plt_bar_stack(
    column=Stages, position = 'stack', 
    labels = c("Mountain stage", "Plain stage", "Time trial"),
    palette = pal_stages, width = 60 #, trim=TRUE
  )

tab3

Rider	Country	Titles	Timeline	Mountain stage\|\|Plain stage\|\|Time trial
Most sucessful riders in the Tour de France
Bernard Hinault The Badger
Eddy Merckx The Cannibal
Jacques Anquetil Maître Jacques
Miguel Indurain Miguelón
Chris Froome Froomey
Greg LeMond The American
Louison Bobet Zonzon
Philippe Thys The Basset Hound

Before you leave, you will find below a slightly more personnal version of this table, with a customized theme. In addition, I added a few words of context in the subtitle, a footnote for additional comments and a source note to details the source of the data. Full code for this version is available here.

Rider	Country	Number of titles	Titles timeline	Stages won
Les forçats de la route
Les forçats de la route, translated as Convicts on the road, is a report by Albert Londres about the Tour de France 1924, an annual men's multiple-stage bicycle contest. In this race across France, the leader is designated with the yellow jersey. The first race was organized in 1903 and in 108 editions, only eight riders have won three or more titles.¹
Rider	Country	Number of titles	Titles timeline	Mountain stage\|\|Plain stage\|\|Time trial
Bernard Hinault The Badger
Eddy Merckx The Cannibal
Jacques Anquetil Maître Jacques
Miguel Indurain Miguelón
Chris Froome Froomey
Greg LeMond The American
Louison Bobet Zonzon
Philippe Thys The Basset Hound
Data: Alastair Rushworth & TidyTuesday \| Table: @BjnNowak
¹ Race not contested from 1915 to 1918 and 1940 to 1946 due to World Wars. Lance Armstrong's wins from 1999 to 2005 were removed due to drugs use, with no alternative winners for those years.

References

Iannone R. et al. Introduction to creating {gt} Tables
Mock T. {gtExtras}