Code to recreate prediction game plot from https://twitter.com/mrcaseb/status/1493242514270261250

prediction_game.R

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

preds <- nflreadr::csv_from_url("https://raw.githubusercontent.com/nflverse/nfldata/master/data/predictions.csv")

g <- nflreadr::load_schedules(2021)

points <- preds |> 
  filter(prediction != 50) |> 
  left_join(g |> select(game_id, week, result), by = "game_id") |> 
  mutate(
    m = case_when(
      week <= 18 ~ 1,
      week == 19 ~ 2,
      week == 20 ~ 3,
      week == 21 ~ 4,
      week == 22 ~ 5
    ),
    r = ifelse(result >= 0, 1, 0),
    points = m * (25 - (100 * (prediction / 100 - r)^2)),
    points = ifelse(result == 0, 0, points)
  ) |> 
  group_by(screen_name) |> 
  summarise(p = sum(points, na.rm = TRUE) |> round(1)) |> 
  ungroup() |> 
  arrange(desc(p)) |> 
  mutate(xaxis = 1:n())

top <- points |> 
  slice_max(p, n = 10) |> 
  mutate(string = glue::glue("#{format(xaxis)} {format(screen_name)} {format(p, nsmall = 1)}"))
rest <- points |> filter(!screen_name %in% c(top$screen_name))
market <- points |> filter(screen_name == "Market")

highlight_col <- "#00685BFF"
market_col <- "#311A92FF"

p <- ggplot(NULL, aes(x = xaxis, y = p)) +
  geom_point(data = rest, alpha = 0.1) +
  geom_point(data = top, alpha = 0.8, color = highlight_col, size = 2) +
  geom_vline(xintercept = market$xaxis, alpha = 0.4, color = market_col, size = 0.5) +
  scale_x_log10("Rank (Logarithmic)") +
  scale_y_continuous("Score", breaks = scales::breaks_pretty(n = 15)) +
  ggthemes::theme_fivethirtyeight(base_size = 11, base_family = "Roboto Condensed") +
  annotate(
    "label", x = market$xaxis, y = -400, label = paste0("#", market$xaxis, " Market ", market$p), 
    color = market_col, family = "Fira Code", fill = "#F0F0F0", size = 2.5
  ) +
  annotate(
    "label", x = 0.5, y = 200, hjust = 0, vjust = 1,
    label = glue::glue("TOP {nrow(top)}:\n", glue::glue_collapse(top$string, sep = "\n")),
    color = highlight_col, family = "Fira Code", fill = "#F0F0F0", size = 2.5
  ) +
  labs(
    title = "NFL Game Data Prediction Game Standings",
    subtitle = "Final Scores of the 2021 Season",
    caption = glue::glue("Figure:@mrcaseb | Data:@LeeSharpeNFL | {lubridate::today()}")
  ) +
  theme(
    plot.title.position = "plot",
    axis.title = element_text(),
    panel.grid.major.x = element_blank()
  ) +
  NULL

ggsave("prediction_game.png", plot = p, width = 16, height = 10, units = "cm", dpi = 600)

Output