nflplotR & the gt Package

Sebastian Carl

The R package {gt} is becoming increasingly popular for creating aesthetically pleasing tables. nflplotR supports rendering of team logos, team wordmarks, and player headshots in gt tables similar to ggplot2. This article will provide some typical examples.

Team Logos & Wordmarks

The functions gt_nfl_logos() and gt_nfl_wordmarks() come with a powerful locations argument that allows usage of gt selection helpers. We will create an example dataframe to show how this all works.

df <- data.frame(
  row_group_column = c("AFC", "NFC", "AFC", "NFC"),
  row_name_column = c("LAC", "SEA"),
  column_a = 11:12,
  column_b = c("KC", "LA")
) 

Our example dataframe in a gt table without any formatting.

gt::gt(df)

row_group_column	row_name_column	column_a	column_b
AFC	LAC	11	KC
NFC	SEA	12	LA
AFC	LAC	11	KC
NFC	SEA	12	LA

The column row_group_column is intended to serve as row group variable so let’s apply this.

gt::gt(df, groupname_col = "row_group_column")

row_name_column	column_a	column_b
AFC
LAC	11	KC
LAC	11	KC
NFC
SEA	12	LA
SEA	12	LA

We also would like to render images in the stub, i.e. the rownames so we tell gt about the row_name_column.

example_table <- gt::gt(
  df, 
  groupname_col = "row_group_column", 
  rowname_col = "row_name_column"
) |> 
  # align the complete table left
  gt::tab_options(
    table.align = "left"
  )
example_table

	column_a	column_b
AFC
LAC	11	KC
LAC	11	KC
NFC
SEA	12	LA
SEA	12	LA

This is our final table. We have valid NFL abbreviations in the cell body, in row group labels and in the stub. We can now use nflplotR to render images instead of those abbreviations.

Cell Body

To render images in the cell body, i.e. the rows of the table, we can either use the columns argument or the appropriate locations helper.

example_table |> 
  nflplotR::gt_nfl_logos(columns = "column_b")

	column_a	column_b
AFC
LAC	11
LAC	11
NFC
SEA	12
SEA	12

Please note, that the locations helper will allow you to selectively apply the function to a set of rows

example_table |> 
  nflplotR::gt_nfl_logos(locations = gt::cells_body(rows = gt::starts_with("LAC")))

	column_a	column_b
AFC
LAC	11
LAC	11
NFC
SEA	12	LA
SEA	12	LA

Row Group Label

Rendering images outside of the cell body will always require the appropriate call to the locations argument. The columns argument cannot handle anything outside cell bodies.

example_table |> 
  nflplotR::gt_nfl_logos(locations = gt::cells_row_groups())

	column_a	column_b
LAC	11	KC
LAC	11	KC
SEA	12	LA
SEA	12	LA

Stub

Now we would like to convert rownames to images.

example_table |> 
  nflplotR::gt_nfl_wordmarks(locations = gt::cells_stub())

	column_a	column_b
AFC
	11	KC
	11	KC
NFC
	12	LA
	12	LA

Column Spanners

We use another table to demonstrate how to convert team names to logos and wordmarks in column spanners.

Here only logos

df <- data.frame(a = 3, b = 4, c = 5, d = 6)

df |> 
  gt::gt() |> 
  gt::tab_spanner("KC", c(a, b)) |> 
  gt::tab_spanner("LAC", c(c, d)) |> 
  nflplotR::gt_nfl_logos(locations = gt::cells_column_spanners())

a	b	c	d
3	4	5	6

And now mix logo and wordmark

df <- data.frame(a = 3, b = 4, c = 5, d = 6)

df |> 
  gt::gt() |> 
  gt::tab_spanner("KC", c(a, b)) |> 
  gt::tab_spanner("LAC", c(c, d)) |> 
  nflplotR::gt_nfl_logos(locations = gt::cells_column_spanners("KC")) |> 
  nflplotR::gt_nfl_wordmarks(locations = gt::cells_column_spanners("LAC"))

a	b	c	d
3	4	5	6

Combine all together

The locations argument allows multiple locations in one call by wrapping them in a list.

example_table |> 
  nflplotR::gt_nfl_wordmarks(locations = gt::cells_stub()) |> 
  nflplotR::gt_nfl_logos(
    locations = list(
      gt::cells_body(), gt::cells_row_groups()
    )
  )

	column_a	column_b
	11
	11
	12
	12

How about Column Labels?

Well…it’s complicated, because gt behaves inconsistent in my opinion.

The actually correct way would be a call to nflplotR::gt_nfl_logos or nflplotR::gt_nfl_wordmarks with the locations argument set to gt::cells_column_labels(). Currently, this wouldn’t render any images in column labels as discussed in the above linked issue.

However, as a convenient workaround, nflplotR supports logos and wordmarks in column labels through gt_nfl_cols_label().

LOGOS:

teams <- nflplotR::valid_team_names() |> head(6)
df <- cbind(1, 2, 3, 4, 5, 6) |> 
  as.data.frame() |> 
  rlang::set_names(teams)

gt::gt(df) |> 
  nflplotR::gt_nfl_cols_label() |> 
  # align the complete table left
  gt::tab_options(
    table.align = "left"
  )

1	2	3	4	5	6

WORDMARKS (note how non matches remain unchanged):

gt::gt(df) |> 
  nflplotR::gt_nfl_cols_label(type = "wordmark") |> 
  # align the complete table left
  gt::tab_options(
    table.align = "left"
  )

AFC
1	2	3	4	5	6

HEADSHOTS:

headshot_df <- data.frame(
  "00-0036355" = 1, 
  "00-0033873" = 2,
  check.names = FALSE
)
gt::gt(headshot_df) |> 
  nflplotR::gt_nfl_cols_label(type = "headshot") |> 
  # align the complete table left
  gt::tab_options(
    table.align = "left"
  )

1	2

Logos and Wordmarks Rendered by nflplotR

This example creates a table that renders all team logos and wordmarks. We split the table into 2 x 16 rows to avoid an overly long table and convert all variables starting with “logo” to logos and all variables starting with “wordmark” to wordmarks.

teams <- nflplotR::valid_team_names()
# remove conference logos from this example
teams <- teams[!teams %in% c("AFC", "NFC", "NFL")]
# create dataframe with all 32 team names
df <- data.frame(
  team_a = head(teams, 16),
  logo_a = head(teams, 16),
  wordmark_a = head(teams, 16),
  team_b = tail(teams, 16),
  logo_b = tail(teams, 16),
  wordmark_b = tail(teams, 16)
)
# create gt table and translate team names to logo/wordmark images
df |>
  gt::gt() |>
  nflplotR::gt_nfl_logos(columns = gt::starts_with("logo")) |>
  nflplotR::gt_nfl_wordmarks(columns = gt::starts_with("wordmark")) |> 
  # align the complete table left
  gt::tab_options(
    table.align = "left"
  )

team_a	logo_a	wordmark_a	team_b	logo_b	wordmark_b
ARI			LA
ATL			LAC
BAL			LV
BUF			MIA
CAR			MIN
CHI			NE
CIN			NO
CLE			NYG
DAL			NYJ
DEN			PHI
DET			PIT
GB			SEA
HOU			SF
IND			TB
JAX			TEN
KC			WAS

Player Headshots

All of the above applies to gt_nfl_headshots() as well. All you need is a gsis ID.

df <- data.frame(
  A = c("00-0036355", "00-0033873"),
  B = c("00-0033077", "00-0035228")
)

df |>
  gt::gt() |> 
  nflplotR::gt_nfl_headshots(columns = gt::everything(), height = 50) |> 
  # align the complete table left
  gt::tab_options(
    table.align = "left"
  )

A	B