How to: Comet plots

F5Days For Everyone

Apr 16, 2021

Hey everyone,

I’m going to try something different and make this week’s walkthrough a freebie — available to everyone and not just paid subscribers. If you know someone who might be interested in learning R while working with NBA data feel free to share this post with them.

If you don’t care about R at all, then I would just delete this email.

Either way, email clients tend to truncate my tutorials so click the title of this post at the top to view it in browser.

In this tutorial, we’re going to focus on the data scraping, processing, and creation of the two charts below. Each step is pretty straight forward, but there are a lot of them so it can end up looking more intimating than it is in reality.

The first plot was inspired by this beautiful graphic by Dr. Cédric Scherer. The code Dr. Scherer used to create this graphic can be found here. One of the differences between my plot and Dr. Scherer’s is that I used the {magick} package to create a custom legend/key, which I’ll show you how to do in this post.

The second plot is riff on a chart Nathan Walker made last year.

Github gist at the end of this post for those that just want the code.

First, lets load our R packages and set up a custom theme.

# Load packages
library(tidyverse)
library(nbastatR)
library(extrafont)
library(ballr)
library(rvest)
library(janitor)
library(hablar)
library(ggforce)
library(ggbrace)
library(magick)
library(ggtext)

# Custom theme
theme_owen <- function () {
  theme_minimal(base_size=9, base_family="Consolas") %+replace%
    theme(
      panel.grid.minor = element_blank(),
      plot.background = element_rect(fill = 'floralwhite', color = "floralwhite")
    )
}

To start off, I’m going to load in the full and abbreviated names for each NBA team. We’ll need them later for joining datasets together later on down the line.

# Get Basketball Reference team name info 
bref_tms <- dictionary_bref_teams()

bref_tms <- bref_tms %>% 
  filter(seasonLast == 2021) %>% 
  select(nameTeamBREF, slugTeamBREF)

bref_tms <- bref_tms %>% 
  mutate(slugTeamBREF = case_when(
    slugTeamBREF == "NJN" ~ "BRK", 
    slugTeamBREF == "CHA" ~ "CHO", 
    slugTeamBREF == "NOH" ~ "NOP",
    TRUE ~ slugTeamBREF
  ))

Our first chart is going to be limited to the highest usage player on each team. I’m going to use the {ballr} package to get every player’s usage rate and then select the highest usage player on each team, with some exceptions.

# Get advanced basketball stats from basketball reference
df_advanced_stats <- NBAPerGameAdvStatistics(season = 2021)

# Select each team's highest usage player (min. 1000 minutes). 
# Exclude LaMelo Ball since he's injured
df_top_usage <- df_advanced_stats %>% 
  filter(mp >= 1000 & tm != "TOT" & player != "LaMelo Ball") %>% 
  group_by(tm) %>% 
  arrange(desc(usgpercent)) %>% 
  slice(1:1) %>% 
  select(player, tm, mp, usgpercent)

Next, we’re going to get each team’s net rating from basektabll-reference.com with the {rvest} package. We’ll also clean the data up and then merge it with our basketball-reference team names that we loaded earlier.

# Get each team's Net Rating from basketball-reference
url <- "https://www.basketball-reference.com/leagues/NBA_2021.html"

bref_tables <- url %>% 
  read_html() %>% 
  html_table()

df_tm_net_rating <- bref_tables[[11]]

df_tm_net_rating <- df_tm_net_rating %>% 
  row_to_names(row_number = 1) %>% 
  clean_names()

df_tm_net_rating <- df_tm_net_rating %>% select(team, n_rtg)

# Join with bref_tms to get slug team names
df_tm_net_rating <- left_join(df_tm_net_rating, bref_tms, by = c("team" = "nameTeamBREF"))

Now, we need to get each player’s individual net rating when they’re on the court. We’ll use the {rvest} package for that as well.

# Get Player On / Off data
url <- "https://www.basketball-reference.com/leagues/NBA_2021_play-by-play.html"

df_player_on_off <- url %>% 
  read_html() %>% 
  html_table(fill = TRUE) %>% 
  as.data.frame()

df_player_on_off <- df_player_on_off %>% 
  row_to_names(row_number = 1) %>% 
  clean_names()

df_player_on_off <- df_player_on_off %>% 
  select(player, tm, mp, on_court, on_off)

# Merge player on/off with team net ratings
df <- left_join(df_player_on_off, df_tm_net_rating, by = c("tm" = "slugTeamBREF"))

Now, we’re going to merge our individual player data with our team data and clean things up a bit.

# Merge player on/off with team net ratings
df <- left_join(df_player_on_off, df_tm_net_rating, by = c("tm" = "slugTeamBREF"))

# Filter out bad data
df <- df %>% filter(tm != "Tm" & tm != "TOT") 

# Change character data to numeric 
df <- df %>% retype()

# Clean up player name
df <- df %>% 
  mutate(player = gsub(" Jr.", "", player), 
         player = gsub(" Sr.", "", player), 
         player = gsub(" III", "", player), 
         player = gsub(" II", "", player), 
         player = gsub(" IV", "", player))

Next, we need to calculate a team’s net rating when the corresponding player is off the court.

# Calculate a team's net rating when a player is OFF the court
df <- df %>% 
  mutate(off_court = case_when(
    on_court >= 0 & on_off >= 0 ~ on_court - on_off, 
    on_court >= 0 & on_off < 0 ~ on_court + abs(on_off), 
    on_court < 0 & on_off >= 0 ~ on_court - on_off, 
    on_court < 0 & on_off < 0 ~ on_court + abs(on_off), 
    TRUE ~ 0
  ))

Okay, our data is in good shape and we’re just about ready to make our plot. But I need to make a few more transformations.

First, I’m going to take our data on the highest usage player on each team and join it with our player on/off and net ratings data.

# Combine usage data with on/off data 
df_comet <- df_top_usage %>% 
  left_join(., df, by = c("player", "tm")) %>% 
  mutate(pos_neg = case_when(
    on_off >= 0 ~ "Better On", 
    TRUE ~ "Worse on",
  ))

I also want to create a new column that combines the player and abbreviated team names. We’ll use those for our labels in our chart.

# Combine player and slug team names together 
df_comet$player_tm <- paste0(df_comet$player, " - ", df_comet$tm)

Okay, now we can make our first chart.

We’re going to use the {ggforce} package to call geom_link which is what make our little comet like shapes. And then we’re going to add a geom_point to each player’s On Court value. We’ll color things green or red depending whether the team is better or worse when said player is on the court.

Everything else is just cosmetics.

comet_plot <- df_comet %>% 
  ggplot() + 
  geom_link(aes(x = off_court, y = fct_reorder(player_tm, on_off), xend = on_court, yend = fct_reorder(player_tm, on_off), color = pos_neg, size = stat(index))) +
  scale_color_manual(values = c("#00A087FF", "#E64B35FF")) +
  scale_size(range = c(.01, 4)) + 
  scale_x_continuous(labels = c("-10", "-5", "0", "+5", "+10"), breaks = seq(-10, 10, 5)) +
  theme_owen() +
  geom_point(
    data = filter(df_comet, on_off > 0),
    aes(on_court, y = fct_reorder(player_tm, on_off), color = pos_neg),
    shape = 21,
    fill = "white",
    size = 3.5
  )  +
  geom_point(
    data = filter(df_comet, on_off < 0),
    aes(on_court, y = fct_reorder(player_tm, on_off), color = pos_neg),
    shape = 21,
    fill = "white",
    size = 3.5
  ) + 
  annotate(geom = 'label', x = 7.5, y = 3.5, label = "Team is worse\nwith them On", family = "Consolas", color = "#E64B35FF", fontface = 'bold', fill = "floralwhite", label.size = 0, size = 3) +
  annotate(geom = 'label', x = -7.5, y = 22.5, label = "Team is better\nwith them On", family = "Consolas", color = "#00A087FF", fontface = 'bold', fill = "floralwhite", label.size = 0, size = 3) +
  theme(legend.position = 'none', 
        plot.title.position = 'plot', 
        axis.text.y = element_text(size = 6),
        plot.title = element_text(face = 'bold', size = 15), 
        plot.subtitle = element_text(size = 7), 
        plot.margin = margin(10, 10, 20, 10)) +
  labs(x = "Net Rating With Player On Or Off The Court", 
       y = "", 
       title = "Net Rating When Player Is On Or Off The Court", 
       subtitle = "Among each team's highest usage player (min. 1000 minutes) | Players sorted by their On/Off Differential") 

# Save Plot
ggsave("Comet_Plot.png", comet_plot, w = 6, h = 6, dpi = 300, type = "cairo")

To add our legend, we’re just going create a smaller version of the same chart, without any labels.

# Make an inset plot that we'll use for our key
inset_plot <- df_comet %>% 
  ggplot() + 
  geom_link(aes(x = 0, y = 1, xend = 1, yend = 1, size = stat(index)), color = "#00A087FF")  + 
  scale_size(range = c(.00001, 3)) + 
  scale_y_continuous(limits = c(.95, 1.05)) +
  scale_x_continuous(limits = c(-.2, 1.2)) +
  theme_owen() +
  labs(title = "KEY") +
  coord_cartesian(clip = 'off') +
  geom_point(aes(x = 1, y = 1), color = "#00A087FF",shape = 21,fill = "white", size = 2.5) + 
  theme(legend.position = 'none', 
        axis.text = element_blank(), 
        axis.title = element_blank(), 
        panel.grid.major = element_blank(), 
        plot.title = element_text(hjust = .5, face = 'bold'), 
        plot.background = element_rect(fill = 'floralwhite', color = "black")) + 
  geom_brace(0.005,1.0175,1.005,1.0175) + 
  annotate(geom = 'text', x = 0.5, y = 1.025, label = "On/Off Differential", family = "Consolas", size = 2.5, hjust = .5, lineheight = 1)  + 
  annotate(geom = 'text', x = 0, y = .985, label = "Net Rating\nWith Player OFF", family = "Consolas", size = 1.85, hjust = .5, lineheight = 1) + 
  annotate(geom = 'text', x = 1, y = .985, label = "Net Rating\nWith Player ON", family = "Consolas", size = 1.85, hjust = .5, lineheight = 1)

# Save our inset plot -- we'll use this later
ggsave("Inset.png",  inset_plot, w = 1.5, h = 1.5, dpi = 300, type = "cairo")

Lastly, we’re going to the use {magick} package to combine these two charts and layer them on top of one another.

# Read in Inset plot
inset <- image_read("Inset.png")

# Read in Comet plot
graf <- image_read("Comet_Plot.png")

# Layer Inset plot on top of Comet plot
image_composite(graf, inset, offset = "+1300+900") %>% image_write("Comet_Plot.png")

Voila!

To make our other chart, we’re going get a dataframe of just team net ratings and then create a couple of dummy data frames that will allow us to annotate our faceted plot

team_nrtg <- df %>% select(n_rtg) %>% distinct()

on_text <- data.frame(value = NA, y = NA, lab = NA,
                           name = factor("Player On Court",levels = c("Player Off Court","Player On Court")))

off_text <- data.frame(value = NA, y = NA, lab = NA,
                      name = factor("Player Off Court",levels = c("Player Off Court","Player On Court")))

Now we can plot our chart.

df %>% 
  filter(mp >= 1000) %>% 
  select(player, tm, off_court, on_court) %>% 
  pivot_longer(-c(player, tm)) %>%
  mutate(name = case_when(
    name == "on_court" ~ "Player On Court",
    name == "off_court" ~ "Player Off Court", 
    TRUE ~ "Net Rating"
  )) %>% 
  ggplot(aes(x = value)) + 
  geom_histogram(alpha = .333, color = '#7570B3FF', fill = '#7570B3FF', bins = 30) + 
  geom_histogram(data = team_nrtg, aes(x = n_rtg, fill = n_rtg), alpha=0.85, color = 'black', fill = '#1B9E77FF', bins = 30) +
  geom_label(data = on_text, aes(x = 10, y = 21.5), label = "Team\nNet Ratings", size = 3, fill = "floralwhite", fontface = 'bold', label.size = 0, color = "#1B9E77FF", family = "Consolas", hjust = .5, alpha = .5) + 
  geom_curve(data = on_text, aes(x = 10, y = 20.5, xend = 2.5, yend = 3), color = "#1B9E77FF", curvature = -0.2, arrow = arrow(length = unit(0.03, "npc"))) + 
  geom_label(data = on_text, aes(x = -10, y = 20), label = "Player ON\nNet Ratings",  size = 3, fill = "floralwhite", fontface = 'bold', label.size = 0, color = "#7570B3FF", family = "Consolas", hjust = .5, alpha = .5) + 
  geom_curve(data = on_text, aes(x = -10, y = 19, xend = -3.75, yend = 15), color = "#7570B3FF", curvature = 0.2, arrow = arrow(length = unit(0.03, "npc"))) + 
  geom_label(data = off_text, aes(x = 10, y = 21.5), label = "Team\nNet Ratings",  size = 3, fill = "floralwhite", fontface = 'bold', label.size = 0, color = "#1B9E77FF", family = "Consolas", hjust = .5, alpha = .5) + 
  geom_curve(data = off_text, aes(x = 10, y = 20.5, xend = 2.5, yend = 3), color = "#1B9E77FF", curvature = -0.2,  arrow = arrow(length = unit(0.03, "npc"))) + 
  geom_label(data = off_text, aes(x = -14.5, y = 20), label = "Player OFF\nNet Ratings", size = 3, fill = "floralwhite", fontface = 'bold', label.size = 0, color = "#7570B3FF", family = "Consolas", hjust = .5, alpha = .5) + 
  geom_curve(data = off_text, aes(x = -14.5, y = 19, xend = -8, yend = 15), color = "#7570B3FF", curvature = 0.2, arrow = arrow(length = unit(0.03, "npc"))) + 
  facet_wrap(~fct_rev(name), nrow = 1) + 
  theme_owen() + 
  scale_x_continuous(limits = c(-20, 20), breaks = seq(-20, 20, 10), labels = c("-20", "-10", "0", "+10", "+20")) + 
  theme(plot.title = element_markdown(size = 11, face = 'bold'), 
        plot.title.position = 'plot', 
        strip.text = element_text(size = 10), 
        panel.spacing.x = unit(2, 'lines'), 
        plot.margin = margin(10, 10, 20, 10), 
        plot.caption = element_text(size = 6.5, vjust = -3, hjust = 1.03)) +
  labs(x = "Net Rating", 
       y = "Count", 
       title = "Distribution Of <span style='color:#7570B3FF'>**Individual Player Net Ratings**</span> vs. <span style='color:#1B9E77FF'>**Team Net Ratings**</span>", 
       subtitle = "Minimum 1000 minutes played", 
       caption = "Chart inspired by Nathan Walker (@bbstats)")


ggsave("Player_vs_Team_Net_Ratings.png", w = 6, h = 6, dpi = 300, type = 'cairo')

        
# Load packages
library(tidyverse)
library(nbastatR)
library(extrafont)
library(ballr)
library(rvest)
library(janitor)
library(hablar)
library(ggforce)
library(ggbrace)
library(magick)
library(ggtext)

# Custom theme
theme_owen <- function () {
  theme_minimal(base_size=9, base_family="Consolas") %+replace%
    theme(
      panel.grid.minor = element_blank(),
      plot.background = element_rect(fill = 'floralwhite', color = "floralwhite")
    )
}

# Get Basketball Reference team name info (for merging purposes later)
bref_tms <- dictionary_bref_teams()
bref_tms <- bref_tms %>% filter(seasonLast == 2021) %>% select(nameTeamBREF, slugTeamBREF)
bref_tms <- bref_tms %>% 
  mutate(slugTeamBREF = case_when(
    slugTeamBREF == "NJN" ~ "BRK", 
    slugTeamBREF == "CHA" ~ "CHO", 
    slugTeamBREF == "NOH" ~ "NOP",
    TRUE ~ slugTeamBREF
  ))

# Get advanced basketabll stats from basketabll reference
df_advanced_stats <- NBAPerGameAdvStatistics(season = 2021)

# Select each team's highest usasge player (min. 1000 minutes). 
# Exclude LaMelo Ball since he's injured
df_top_usage <- df_advanced_stats %>% 
  filter(mp >= 1000 & tm != "TOT" & player != "LaMelo Ball") %>% 
  group_by(tm) %>% 
  arrange(desc(usgpercent)) %>% 
  slice(1:1) %>% 
  select(player, tm, mp, usgpercent)

# Get each team's Net Rating from basketabll-reference
url <- "https://www.basketball-reference.com/leagues/NBA_2021.html"

bref_tables <- url %>% 
  read_html() %>% 
  html_table()

df_tm_net_rating <- bref_tables[[11]]

df_tm_net_rating <- df_tm_net_rating %>% 
  row_to_names(row_number = 1) %>% 
  clean_names()

df_tm_net_rating <- df_tm_net_rating %>% select(team, n_rtg)

# Join with bref_tms to get slug team names
df_tm_net_rating <- left_join(df_tm_net_rating, bref_tms, by = c("team" = "nameTeamBREF"))

# Get Player On / Off data
url <- "https://www.basketball-reference.com/leagues/NBA_2021_play-by-play.html"

df_player_on_off <- url %>% 
  read_html() %>% 
  html_table(fill = TRUE) %>% 
  as.data.frame()

df_player_on_off <- df_player_on_off %>% 
  row_to_names(row_number = 1) %>% 
  clean_names()

df_player_on_off <- df_player_on_off %>% 
  select(player, tm, mp, on_court, on_off)

# Merge player on/off with team net ratings
df <- left_join(df_player_on_off, df_tm_net_rating, by = c("tm" = "slugTeamBREF"))

# Filter out bad data
df <- df %>% filter(tm != "Tm" & tm != "TOT") 

# Change character data to numeric 
df <- df %>% retype()

# Clean up player name
df <- df %>% 
  mutate(player = gsub(" Jr.", "", player), 
         player = gsub(" Sr.", "", player), 
         player = gsub(" III", "", player), 
         player = gsub(" II", "", player), 
         player = gsub(" IV", "", player)) 

# Calcualte a team's net rating when a player is OFF the court
df <- df %>% 
  mutate(off_court = case_when(
    on_court >= 0 & on_off >= 0 ~ on_court - on_off, 
    on_court >= 0 & on_off < 0 ~ on_court + abs(on_off), 
    on_court < 0 & on_off >= 0 ~ on_court - on_off, 
    on_court < 0 & on_off < 0 ~ on_court + abs(on_off), 
    TRUE ~ 0
  ))

# Comet Plot --------------------------------------------------------------

# Use our dataframe of the top usage player on each team and find out whether 
# their team is better or wose with them on/off the court

df_comet <- df_top_usage %>% 
  left_join(., df, by = c("player", "tm")) %>% 
  mutate(pos_neg = case_when(
    on_off >= 0 ~ "Better On", 
    TRUE ~ "Worse on",
  )) 

# Combine player and slug team names together 
df_comet$player_tm <- paste0(df_comet$player, " - ", df_comet$tm)

comet_plot <- df_comet %>% 
  ggplot() + 
  geom_link(aes(x = off_court, y = fct_reorder(player_tm, on_off), xend = on_court, yend = fct_reorder(player_tm, on_off), color = pos_neg, size = stat(index))) +
  scale_color_manual(values = c("#00A087FF", "#E64B35FF")) +
  scale_size(range = c(.01, 4)) + 
  scale_x_continuous(labels = c("-10", "-5", "0", "+5", "+10"), breaks = seq(-10, 10, 5)) +
  theme_owen() +
  geom_point(
    data = filter(df_comet, on_off > 0),
    aes(on_court, y = fct_reorder(player_tm, on_off), color = pos_neg),
    shape = 21,
    fill = "white",
    size = 3.5
  )  +
  geom_point(
    data = filter(df_comet, on_off < 0),
    aes(on_court, y = fct_reorder(player_tm, on_off), color = pos_neg),
    shape = 21,
    fill = "white",
    size = 3.5
  ) + 
  annotate(geom = 'label', x = 7.5, y = 3.5, label = "Team is worse\nwith them On", family = "Consolas", color = "#E64B35FF", fontface = 'bold', fill = "floralwhite", label.size = 0, size = 3) +
  annotate(geom = 'label', x = -7.5, y = 22.5, label = "Team is better\nwith them On", family = "Consolas", color = "#00A087FF", fontface = 'bold', fill = "floralwhite", label.size = 0, size = 3) +
  theme(legend.position = 'none', 
        plot.title.position = 'plot', 
        axis.text.y = element_text(size = 6),
        plot.title = element_text(face = 'bold', size = 15), 
        plot.subtitle = element_text(size = 7), 
        plot.margin = margin(10, 10, 20, 10)) +
  labs(x = "Net Rating With Player On Or Off The Court", 
       y = "", 
       title = "Net Rating When Player Is On Or Off The Court", 
       subtitle = "Among each team's highest usage player (min. 1000 minutes) | Players sorted by their On/Off Differential") 

# Save Plot
ggsave("Comet_Plot.png", comet_plot, w = 6, h = 6, dpi = 300, type = "cairo")

# Make an inset plot that we'll use for our key
inset_plot <- df_comet %>% 
  ggplot() + 
  geom_link(aes(x = 0, y = 1, xend = 1, yend = 1, size = stat(index)), color = "#00A087FF")  + 
  scale_size(range = c(.00001, 3)) + 
  scale_y_continuous(limits = c(.95, 1.05)) +
  scale_x_continuous(limits = c(-.2, 1.2)) +
  theme_owen() +
  labs(title = "KEY") +
  coord_cartesian(clip = 'off') +
  geom_point(aes(x = 1, y = 1), color = "#00A087FF",shape = 21,fill = "white", size = 2.5) + 
  theme(legend.position = 'none', 
        axis.text = element_blank(), 
        axis.title = element_blank(), 
        panel.grid.major = element_blank(), 
        plot.title = element_text(hjust = .5, face = 'bold'), 
        plot.background = element_rect(fill = 'floralwhite', color = "black")) + 
  geom_brace(0.005,1.0175,1.005,1.0175) + 
  annotate(geom = 'text', x = 0.5, y = 1.025, label = "On/Off Differential", family = "Consolas", size = 2.5, hjust = .5, lineheight = 1)  + 
  annotate(geom = 'text', x = 0, y = .985, label = "Net Rating\nWith Player OFF", family = "Consolas", size = 1.85, hjust = .5, lineheight = 1) + 
  annotate(geom = 'text', x = 1, y = .985, label = "Net Rating\nWith Player ON", family = "Consolas", size = 1.85, hjust = .5, lineheight = 1)

# Save our inset plot -- we'll use this later
ggsave("Inset.png",  inset_plot, w = 1.5, h = 1.5, dpi = 300, type = "cairo")

# Read in Inset plot
inset <- image_read("Inset.png")

# Read in Comet plot
graf <- image_read("Comet_Plot.png")

# Layer Inset plot on top of Comet plot
image_composite(graf, inset, offset = "+1300+900") %>% image_write("Comet_Plot.png")

# Histogram plot ----------------------------------------------------------

team_nrtg <- df %>% select(n_rtg) %>% distinct()

on_text <- data.frame(value = NA, y = NA, lab = NA,
                           name = factor("Player On Court",levels = c("Player Off Court","Player On Court")))

off_text <- data.frame(value = NA, y = NA, lab = NA,
                      name = factor("Player Off Court",levels = c("Player Off Court","Player On Court")))

df %>% 
  filter(mp >= 1000) %>% 
  select(player, tm, off_court, on_court) %>% 
  pivot_longer(-c(player, tm)) %>%
  mutate(name = case_when(
    name == "on_court" ~ "Player On Court",
    name == "off_court" ~ "Player Off Court", 
    TRUE ~ "Net Rating"
  )) %>% 
  ggplot(aes(x = value)) + 
  geom_histogram(alpha = .333, color = '#7570B3FF', fill = '#7570B3FF', bins = 30) + 
  geom_histogram(data = team_nrtg, aes(x = n_rtg, fill = n_rtg), alpha=0.85, color = 'black', fill = '#1B9E77FF', bins = 30) +
  geom_label(data = on_text, aes(x = 10, y = 21.5), label = "Team\nNet Ratings", size = 3, fill = "floralwhite", fontface = 'bold', label.size = 0, color = "#1B9E77FF", family = "Consolas", hjust = .5, alpha = .5) + 
  geom_curve(data = on_text, aes(x = 10, y = 20.5, xend = 2.5, yend = 3), color = "#1B9E77FF", curvature = -0.2, arrow = arrow(length = unit(0.03, "npc"))) + 
  geom_label(data = on_text, aes(x = -10, y = 20), label = "Player ON\nNet Ratings",  size = 3, fill = "floralwhite", fontface = 'bold', label.size = 0, color = "#7570B3FF", family = "Consolas", hjust = .5, alpha = .5) + 
  geom_curve(data = on_text, aes(x = -10, y = 19, xend = -3.75, yend = 15), color = "#7570B3FF", curvature = 0.2, arrow = arrow(length = unit(0.03, "npc"))) + 
  geom_label(data = off_text, aes(x = 10, y = 21.5), label = "Team\nNet Ratings",  size = 3, fill = "floralwhite", fontface = 'bold', label.size = 0, color = "#1B9E77FF", family = "Consolas", hjust = .5, alpha = .5) + 
  geom_curve(data = off_text, aes(x = 10, y = 20.5, xend = 2.5, yend = 3), color = "#1B9E77FF", curvature = -0.2,  arrow = arrow(length = unit(0.03, "npc"))) + 
  geom_label(data = off_text, aes(x = -14.5, y = 20), label = "Player OFF\nNet Ratings", size = 3, fill = "floralwhite", fontface = 'bold', label.size = 0, color = "#7570B3FF", family = "Consolas", hjust = .5, alpha = .5) + 
  geom_curve(data = off_text, aes(x = -14.5, y = 19, xend = -8, yend = 15), color = "#7570B3FF", curvature = 0.2, arrow = arrow(length = unit(0.03, "npc"))) + 
  facet_wrap(~fct_rev(name), nrow = 1) + 
  theme_owen() + 
  scale_x_continuous(limits = c(-20, 20), breaks = seq(-20, 20, 10), labels = c("-20", "-10", "0", "+10", "+20")) + 
  theme(plot.title = element_markdown(size = 11, face = 'bold'), 
        plot.title.position = 'plot', 
        strip.text = element_text(size = 10), 
        panel.spacing.x = unit(2, 'lines'), 
        plot.margin = margin(10, 10, 20, 10), 
        plot.caption = element_text(size = 6.5, vjust = -3, hjust = 1.03)) +
  labs(x = "Net Rating", 
       y = "Count", 
       title = "Distribution Of <span style='color:#7570B3FF'>**Individual Player Net Ratings**</span> vs. <span style='color:#1B9E77FF'>**Team Net Ratings**</span>", 
       subtitle = "Minimum 1000 minutes played")

ggsave("Player_vs_Team_Net_Ratings.png", w = 6, h = 6, dpi = 300, type = 'cairo')