Skip to contents

NFL EPA Scatter

Saiem Gilani
@saiemgilani @saiemgilani

Source: vignettes/recipe-nfl-epa-scatter.Rmd
recipe-nfl-epa-scatter.Rmd
On this page

The brief: the week after the Super Bowl, the social team wants the season’s offense-vs-defense chart for X and Bluesky (1200 x 675) and Instagram (1080 x 1080), and it should say something about the two teams that just played. This is the chart every NFL analyst makes: each team’s offensive EPA per play against the EPA per play its defense allowed, with logos for points. The data is nflverse play-by-play through nflreadr.

library(sdvplotR)
library(ggplot2)
library(dplyr, warn.conflicts = FALSE)

season <- 2025
out_dir <- tempfile("sdvplotR-recipe-") # where the exports go; use your own folder
dir.create(out_dir)

1. Get the data

One season of play-by-play, cut to regular-season runs and passes with an EPA value, then averaged twice: by the team with the ball and by the team on defense. The Super Bowl comes from the schedule, so the chart can point at it.

plays <- nflreadr::load_pbp(season) |>
  filter(season_type == "REG", play_type %in% c("pass", "run"), !is.na(epa))

offense <- plays |>
  group_by(team = posteam) |>
  summarise(off_epa = mean(epa), plays = n())
defense <- plays |>
  group_by(team = defteam) |>
  summarise(def_epa = mean(epa))
stopifnot(is.character(offense$team), is.character(defense$team)) # one type on both sides of the key

teams <- inner_join(offense, defense, by = "team") |>
  mutate(
    off_rank = min_rank(desc(off_epa)),
    def_rank = min_rank(def_epa) # less EPA allowed is better
  )
stopifnot(nrow(teams) == 32)

sb <- nflreadr::load_schedules(season) |> filter(game_type == "SB")
winner <- if (sb$home_score > sb$away_score) sb$home_team else sb$away_team
loser <- setdiff(c(sb$home_team, sb$away_team), winner)
sprintf("Super Bowl: %s %d, %s %d", sb$away_team, sb$away_score, sb$home_team, sb$home_score)
#> [1] "Super Bowl: SEA 29, NE 13"
filter(teams, team %in% c(winner, loser))
#> # A tibble: 2 × 6
#>   team  off_epa plays def_epa off_rank def_rank
#>   <chr>   <dbl> <int>   <dbl>    <int>    <int>
#> 1 NE     0.159   1017 -0.0471        1       11
#> 2 SEA    0.0328   997 -0.116        13        2

The champion had the No. 2 defense and the runner-up the No. 1 offense: that is the story the chart should tell.

2. The first draft

Thirty-two dots and their abbreviations, the five-minute version.

ggplot(teams, aes(off_epa, def_epa)) +
  geom_point() +
  geom_text(aes(label = team), nudge_y = 0.006, size = 3) +
  labs(x = "off_epa", y = "def_epa")

First draft: a plain scatter of 32 NFL teams' offensive EPA per play against defensive EPA per play allowed, labelled with abbreviations that overlap in the middle.

It is all there, but it does not read: labels collide in the middle of the pack, a reader has to work out that low is good on the y axis, and nothing says what the chart is about.

3. Logos for points

geom_sdv_logos() puts each team’s logo on its point, and the nflverse abbreviations resolve as they are. width is a share of the panel’s width, so the logos keep their size when the plot is saved at another size. Logos are drawn inside the panel, so both scales get a little extra room (expansion()) and no logo at an edge is cut off.

ggplot(teams, aes(off_epa, def_epa)) +
  geom_sdv_logos(aes(team = team), league = "nfl", width = 0.055) +
  scale_x_continuous(expand = expansion(mult = 0.07)) +
  scale_y_continuous(expand = expansion(mult = 0.1)) +
  labs(x = "Offense EPA per play", y = "Defense EPA per play allowed")

The same scatter with each NFL team drawn as its logo instead of a point.

4. Point the axes the same way

Readers expect “up and to the right” to be good. scale_y_reverse() puts the best defenses on top; geom_mean_lines() draws dashed league-average lines, and a label in each corner turns the scatter into four quadrants anyone can read. Under a reversed scale -Inf is the top edge, so the corners are placed with Inf and -Inf and the labels sit just inside them.

corners <- data.frame(
  x = c(Inf, -Inf, Inf, -Inf),
  y = c(-Inf, -Inf, Inf, Inf), # reversed scale: -Inf is the top
  label = c(
    "Good offense, good defense", "Bad offense, good defense",
    "Good offense, bad defense", "Bad offense, bad defense"
  ),
  hjust = c(1.02, -0.02, 1.02, -0.02),
  vjust = c(1.5, 1.5, -0.6, -0.6)
)

quadrants <- list(
  geom_mean_lines(aes(x0 = off_epa, y0 = def_epa), color = "grey60", linewidth = 0.4),
  geom_text(
    aes(x, y, label = label, hjust = hjust, vjust = vjust),
    data = corners, inherit.aes = FALSE,
    size = 3, color = "grey55", fontface = "italic"
  ),
  scale_x_continuous(expand = expansion(mult = 0.07)),
  scale_y_reverse(expand = expansion(mult = 0.12))
)

ggplot(teams, aes(off_epa, def_epa)) +
  quadrants +
  geom_sdv_logos(aes(team = team), league = "nfl", width = 0.055) +
  labs(x = "Offense EPA per play", y = "Defense EPA per play allowed")

The logo scatter with the defense axis flipped, dashed lines at the league averages, and a grey italic label in each corner naming the quadrant.

Look for the two finalists. Seattle is half hidden: its numbers are almost the same as Jacksonville’s, and the Jaguars logo covers it. That is one of the two teams the chart is about, and nothing on it points at either of them.

5. Tell the story

Now make it about the Super Bowl. The other 30 logos fade (alpha), and the two finalists are drawn last and a little larger, which also lifts Seattle out from under Jacksonville. Each gets a callout in its team color (scale_color_sdv()), placed beside the logo rather than on it: the champion’s to the right, the runner-up’s to the left, a fixed share of the x range away. The title states the finding instead of describing the axes, the subtitle carries the definitions and the caption the source.

Titles are the part that breaks when one chart goes out at two sizes. ggtext’s element_textbox_simple() wraps the title and subtitle to the width of the canvas, so the same theme works on the wide post and on the square one. Everything goes in one function of the logo size, ready for the exports.

finalists <- c(winner, loser)
gap <- 0.08 * diff(range(teams$off_epa)) # clear of a logo, in EPA
callouts <- teams |>
  filter(team %in% finalists) |>
  mutate(
    note = if_else(
      team == winner,
      paste0("Champion\nNo. ", def_rank, " defense"),
      paste0("Runner-up\nNo. ", off_rank, " offense")
    ),
    side = if_else(team == winner, 1, -1),
    x = off_epa + side * gap,
    hjust = if_else(side > 0, 0, 1)
  )

title <- sprintf(
  "Super Bowl LX matched the No. %d offense against the No. %d defense",
  teams$off_rank[teams$team == loser], teams$def_rank[teams$team == winner]
)
subtitle <- paste0(
  "EPA per play on runs and passes, ", season, " regular season. ",
  "Defense axis flipped: up and right is better. ", winner, " won ",
  max(sb$home_score, sb$away_score), "-", min(sb$home_score, sb$away_score), "."
)

epa_chart <- function(logo_width = 0.055) {
  ggplot(teams, aes(off_epa, def_epa)) +
    quadrants +
    geom_sdv_logos(
      aes(team = team),
      data = filter(teams, !team %in% finalists),
      league = "nfl", width = logo_width, alpha = 0.3
    ) +
    geom_sdv_logos(
      aes(team = team),
      data = filter(teams, team %in% finalists),
      league = "nfl", width = logo_width * 1.3
    ) +
    geom_text(
      aes(x = x, label = note, color = team, hjust = hjust),
      data = callouts, size = 3.2, fontface = "bold", lineheight = 0.9
    ) +
    scale_color_sdv(league = "nfl") +
    labs(
      title = title, subtitle = subtitle,
      x = "Offense EPA per play", y = "Defense EPA per play allowed",
      caption = "Data: nflverse via nflreadr  |  Chart: sdvplotR"
    ) +
    theme_minimal(base_size = 10) +
    theme(
      plot.title.position = "plot",
      plot.title = ggtext::element_textbox_simple(
        face = "bold", size = 14, margin = margin(b = 4)
      ),
      plot.subtitle = ggtext::element_textbox_simple(
        color = "grey40", size = 9.5, margin = margin(b = 8)
      ),
      plot.caption = element_text(color = "grey45", size = 7.5),
      axis.title = element_text(color = "grey40", size = 9),
      axis.text = element_text(color = "grey55"),
      panel.grid.minor = element_blank(),
      plot.margin = margin(12, 14, 8, 12),
      plot.background = element_rect(fill = "white", color = NA)
    )
}

epa_chart()

Final chart: NFL offense EPA against defense EPA allowed, defense axis flipped, with 30 faded logos and the two Super Bowl teams drawn full strength with callouts naming the champion's defense rank and the runner-up's offense rank.

6. Export at social sizes

Social sites resize whatever they get, so export at the size they show: 1200 x 675 for X and Bluesky, 1080 x 1080 for Instagram. ggsave() takes the size in inches and a dpi, and inches times dpi is the pixel count. The square canvas has a narrower panel than the wide one, and width is a share of the panel, so the square gets a slightly larger share to keep the logos the same size on screen.

wide <- file.path(out_dir, "nfl_epa_1200x675.png")
square <- file.path(out_dir, "nfl_epa_1080x1080.png")
ggsave(wide, epa_chart(), width = 8, height = 4.5, dpi = 150)
ggsave(square, epa_chart(logo_width = 0.075), width = 7.2, height = 7.2, dpi = 150)

files <- magick::image_read(c(wide, square))
magick::image_info(files)[c("width", "height")]
#> # A tibble: 2 × 2
#>   width height
#>   <int>  <int>
#> 1  1200    675
#> 2  1080   1080
magick::image_read(wide)

The finished EPA chart exported at 1200 by 675 pixels for X and Bluesky.

The square cut, from the same function:

magick::image_read(square)

The finished EPA chart exported at 1080 by 1080 pixels for Instagram.

Our Authors

Citation

To cite the sdvplotR R package in publications, use:

BibTeX Citation

@misc{gilani_carl_lee_kay_weatherman_sdvplotR,
  author = {Gilani, Saiem and Carl, Sebastian and Lee, Jared and Kay, Camden and Weatherman, Andrew},
  title = {sdvplotR: The SportsDataverse's R Package for Multi-League Sports Plotting.},
  url = {https://sdvplotR.sportsdataverse.org},
  year = {2026}
}