The Simpsons Sweet Spot: Where Ratings and Viewership Meet

This visualization was created using {ggplot2} for plotting, {biscale} for the bivariate color scale, and {ggimage} to display each episode’s image at its corresponding IMDb rating and viewership position. The bivariate color-coded circumferences classify episodes into tertiles along both axes, blending blue, red, and violet to highlight different rating-viewership combinations. Data wrangling and processing were done in R, making use of the {tidyverse} ecosystem.

About the Data

The Simpsons Dataset provides a comprehensive look into the iconic animated series, covering over 600 episodes. Curated by Prashant Banerjee on Kaggle, this dataset includes four key files: characters, locations, episodes, and script lines. Originally, the dataset was scraped by Todd W. Schneider for his analysis, with the scraper made available on GitHub. The dataset has since been rehosted and refined for broader accessibility.

The characters dataset (simpsons_characters.csv) provides details on individual characters, including gender and name normalization. The episodes dataset (simpsons_episodes.csv) captures IMDb ratings, viewership in millions, and air dates, enabling insights into episode popularity. The locations dataset (simpsons_locations.csv) records frequently featured places, while the script lines dataset (simpsons_script_lines.csv) documents spoken words, character dialogues, and word counts.

Figure 1: This graphic visualizes the relationship between IMDb ratings (X-axis) and U.S. viewership in millions (Y-axis) for 151 episodes of The Simpsons. Each episode is represented by its official image, placed according to its rating and viewership, with a bivariate color-coded circumference created using the {biscale} package. The colors reflect a 3×3 tertile classification, blending blue, red, and violet to indicate different combinations of high, medium, and low ratings and viewership. Data was processed and visualized in R using {ggplot2}, {biscale}, and {ggimage} for image integration.

How I made this graphic?

Key Learnings

1. Getting a specialized Simpsons palette from {ggsci} (Xiao 2024)

2. Showing colour palettes using {scales} (Wickham, Pedersen, and Seidel 2023) function scales::show_col()

3. Making bivariate colour scales using {biscale} - inspiration article.

Loading required libraries, data import & creating custom functions.

# Data Import and Wrangling Tools
library(tidyverse)            # All things tidy

# Final plot tools
library(scales)               # Nice Scales for ggplot2
library(fontawesome)          # Icons display in ggplot2
library(ggtext)               # Markdown text support for ggplot2
library(showtext)             # Display fonts in ggplot2
library(colorspace)           # Lighten and Darken colours
library(magick)               # Image manipulation
library(httr)                 # Downloading Google images
library(biscale)              # Bivariate Scales in R
library(patchwork)            # Composing gpglot2 plots

simpsons_episodes <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2025/2025-02-04/simpsons_episodes.csv')

Visualization Parameters

# Font for titles
font_add_google("Rock Salt",
  family = "title_font"
) 

# Font for the caption
font_add_google("Saira Extra Condensed",
  family = "caption_font"
) 

# Font for plot text
font_add_google("Titillium Web",
  family = "body_font"
) 

showtext_auto()

# Get a specialized Simpsons Palette from {ggsci}
library("scales")

mypal <- ggsci::pal_simpsons("springfield")(16)
seecolor::print_color(mypal)

# Using scales::show_col() to display a palette
mypal |> scales::show_col(ncol = 4)

# A base Colour
bg_col <- mypal[1]
seecolor::print_color(bg_col)

# Colour for highlighted text
text_hil <- mypal[13]
seecolor::print_color(text_hil)

# Colour for the text
text_col <- mypal[7]
seecolor::print_color(text_col)

blend_cols <- c(mypal[11], mypal[15])
seecolor::print_color(blend_cols)

# Define Base Text Size
bts <- 90 

# Caption stuff for the plot
sysfonts::font_add(
  family = "Font Awesome 6 Brands",
  regular = here::here("docs", "Font Awesome 6 Brands-Regular-400.otf")
)
github <- "&#xf09b"
github_username <- "aditya-dahiya"
xtwitter <- "&#xe61b"
xtwitter_username <- "@adityadahiyaias"
social_caption_1 <- glue::glue("<span style='font-family:\"Font Awesome 6 Brands\";'>{github};</span> <span style='color: {text_hil}'>{github_username}  </span>")
social_caption_2 <- glue::glue("<span style='font-family:\"Font Awesome 6 Brands\";'>{xtwitter};</span> <span style='color: {text_hil}'>{xtwitter_username}</span>")
plot_caption <- paste0(
  "**Data:** Todd W. Schneider & Prashant Banerjee", 
  " |  **Code:** ", 
  social_caption_1, 
  " |  **Graphics:** ", 
  social_caption_2
  )
rm(github, github_username, xtwitter, 
   xtwitter_username, social_caption_1, 
   social_caption_2)

# Add text to plot-------------------------------------------------
plot_title <- "The Simpsons:\nRatings vs. Viewership"

plot_subtitle <- "While some Simpsons episodes boast sky-high IMDb ratings, they aren’t necessarily the most-watched. Overall, there appears to be little to no correlation between viewership numbers and episode ratings."

Exploratory Data Analysis and Wrangling

# Prepare the data for the bivariate scale
df1 <- simpsons_episodes |> 
  arrange(number_in_season, number_in_series) |> 
  mutate(id = row_number()) |> 
  rename(viewers = us_viewers_in_millions) |> 
  select(id, imdb_rating, viewers, number_in_season, image_url, 
         title, season) |> 
  biscale::bi_class(
    x = imdb_rating, y = viewers,
    dim = 3,
    style = "quantile"
  )

# Creating a bivariate palette: view the palette
biscale::bi_pal("DkViolet", dim = 3)

# Draw a legend as a {ggplot2} object
biscale::bi_legend(
  pal = "DkViolet",
  xlab = "Higher % White ",
  ylab = "Higher Income ",
  size = 12,
  arrows = T
  )


# Draw a legend as a {ggplot2} object
g_legend <- biscale::bi_legend(
  pal = "DkViolet",
  xlab = "Higher Ratings on IMDb",
  ylab = "Higher viewership in USA",
  size = bts * 0.7,
  pad_color = bg_col,
  pad_width = 0.5,
  arrows = T
  ) +
  theme(
    axis.title = element_text(
      family = "caption_font",
      colour = text_col,
      margin = margin(0,0,0,0, "mm")
    ),
    panel.background = element_rect(
      fill = "transparent",
      colour = "transparent"
    ),
    plot.background = element_rect(
      fill = "transparent",
      colour = "transparent"
    )
  )

Downloading images for the selected episodes

# Few outlier episodes to display the photo
selected_ids <- c(60, 61, 57, 69, 65, 59, 55, 
                  39, 36, 134, 146, 5, 102, 
                  127, 115, 1, 41)

df2 <- df1 |> 
  filter(id %in% selected_ids) |> 
  mutate(
    image_path = paste0("data_vizs/temp_simpsons_", id, ".png")
  )

# Get a custom google search engine and API key
# Tutorial: https://developers.google.com/custom-search/v1/overview
# Tutorial 2: https://programmablesearchengine.google.com/

# From:https://developers.google.com/custom-search/v1/overview
# google_api_key <- "LOAD YOUR GOOGLE API KEY HERE"

# From: https://programmablesearchengine.google.com/controlpanel/all
# my_cx <- "GET YOUR CUSTOM SEARCH ENGINE ID HERE"

# Improved function to download and save food images
download_simpsons <- function(i) {
  
  api_key <- google_api_key
  cx <- my_cx
  
  # Build the API request URL with additional filters
  url <- paste0(
    "https://www.googleapis.com/customsearch/v1?q=",
    URLencode(paste0("Simpsons episode ", df2$title[i], " image")),
    "&cx=", cx,
    "&searchType=image",
    "&key=", api_key,
    "&num=1"                 # Fetch only one result
  )
  
  # Make the request
  response <- GET(url)
  if (response$status_code != 200) {
    warning("Failed to fetch data for: ", 
            dfww2$country[i])
    return(NULL)
  }
  
  # Parse the response
  result <- content(response, "parsed")
  
  # Extract the image URL
  if (!is.null(result$items)) {
    image_url <- result$items[[1]]$link
  } else {
    warning("No results found for: ", dfww2$country[i])
    return(NULL)
  }
  
  # Process the image
  im <- magick::image_read(image_url) |> 
    # Crop the image into a circle 
    # (Credits: https://github.com/doehm/cropcircles)
    cropcircles::circle_crop(
      border_colour = "black",
      border_size = 0.1
    ) |>
    image_read() |> 
    image_background(color = bg_col) |> 
    image_resize("x300") |> 
    # Save or display the result
    image_write(
      here::here(
        "data_vizs", 
        paste0("temp_simpsons_", df2$id[i], ".png")
        )
    )
}

# Iterate through each state and download images
for (i in 1:nrow(df2)) {
  download_simpsons(i)
}

The Base Plot

g <- df1 |> 
  ggplot(
    mapping = aes(
      x = imdb_rating,
      y = viewers
    )
  ) +
  geom_point(
    mapping = aes(colour = bi_class),
    position = position_jitter(
      width = 0.1
    ),
    size = 6
  ) +
  
  # Selected episodes
  ggimage::geom_image(
    data = df2,
    mapping = aes(
      image = image_path
    ),
    size = 0.08
  ) +
  geom_point(
    mapping = aes(colour = bi_class),
    data = df2,
    pch = 21,
    size = 30,
    stroke = 4,
    fill = "transparent"
  ) +
  geom_text(
    data = df2,
    mapping = aes(
      label = paste0(
        title,
        "\nSeason ", 
        season,
        ",  Ep.", 
        number_in_season
      )
    ),
    size = 12,
    nudge_y = -0.8,
    lineheight = 0.25,
    family = "caption_font",
    colour = darken(text_hil, 0.5)
  ) +
  biscale::bi_scale_color(
    pal = "DkViolet", 
    dim = 3
  ) +
  scale_y_continuous(
    labels = label_number(suffix = " M"),
    breaks = seq(0, 16, 2)
  ) +
  labs(
    title = plot_title,
    subtitle = str_wrap(plot_subtitle, 80),
    caption = plot_caption,
    x = "Rating of each episode on IMDb",
    y = "Viewership in Millions (USA)"
  ) +
  theme_minimal(
    base_family = "body_font",
    base_size = bts
  ) +
  theme(
    
    # Overall Plot
    plot.margin = margin(5,5,5,5, "mm"),
    plot.title.position = "plot",
    panel.grid = element_line(
      colour = alpha(mypal[5], 0.5),
      linewidth = 0.6,
      linetype = 3
    ),
    axis.line = element_line(
      colour = mypal[5],
      linetype = 1,
      linewidth = 0.9,
      arrow = arrow(angle = 30, length = unit(10, "mm"))
    ),
    axis.ticks = element_line(
      colour = mypal[5],
      linewidth = 0.6
    ),
    axis.ticks.length = unit(2, "mm"),
    text = element_text(
      colour = text_col,
      margin = margin(0,0,0,0, "mm"),
      hjust = 0.5,
      vjust = 0.5,
      lineheight = 0.3
    ),
    
    # Axis Text
    axis.text = element_text(
      margin = margin(0,0,0,0, "mm"),
      colour = text_col
    ),
    axis.title = element_text(
      margin = margin(0.5,0.5,0.5,0.5, "mm"), 
      face = "bold"
    ),
    
    # Labels
    plot.title = element_text(
      colour = text_hil,
      margin = margin(5,0,5,0, "mm"),
      size = bts * 2.5,
      lineheight = 0.3,
      hjust = 0.5,
      family = "title_font"
    ),
    plot.subtitle = element_text(
      colour = text_hil,
      margin = margin(5,0,0,0, "mm"),
      size = bts * 1.2,
      hjust = 0.5,
      lineheight = 0.3,
      family = "body_font"
    ),
    plot.caption = element_textbox(
      family = "caption_font",
      margin = margin(5,0,2,0, "mm"),
      hjust = 0.5,
      colour = text_hil
    ),
    legend.position = "none"
  )

g_full <- g +
  inset_element(
    g_legend,
    left = 0.02, right = 0.35,
    bottom = 0.6, top = 1,
    align_to = "panel"
  ) +
  plot_annotation(
    theme = theme(
    panel.background = element_rect(
      fill = bg_col,
      colour = bg_col
    ),
    plot.background = element_rect(
      fill = bg_col,
      colour = bg_col
      )
    )
  )

ggsave(
  filename = here::here(
    "data_vizs",
    "tidy_simpsons_feb25.png"
  ),
  plot = g_full,
  width = 400,
  height = 500,
  units = "mm",
  bg = bg_col
)

Savings the thumbnail for the webpage

# Saving a thumbnail

library(magick)

# Saving a thumbnail for the webpage
image_read(here::here("data_vizs", 
                      "tidy_simpsons_feb25.png")) |> 
  image_resize(geometry = "x400") |> 
  image_write(
    here::here(
      "data_vizs", 
      "thumbnails", 
      "tidy_simpsons_feb25.png"
    )
  )

# Define the folder path
folder_path <- file.path(getwd(), "data_vizs")

# List all files in the folder that start with "temp_simpsons_"
files_to_delete <- list.files(folder_path, pattern = "^temp_simpsons_", full.names = TRUE)

# Delete the files
file.remove(files_to_delete)

# Print confirmation
cat(length(files_to_delete), "files deleted.\n")

Session Info

# Data Import and Wrangling Tools
library(tidyverse)            # All things tidy

# Final plot tools
library(scales)               # Nice Scales for ggplot2
library(fontawesome)          # Icons display in ggplot2
library(ggtext)               # Markdown text support for ggplot2
library(showtext)             # Display fonts in ggplot2
library(colorspace)           # Lighten and Darken colours
library(magick)               # Image manipulation
library(httr)                 # Downloading Google images
library(biscale)              # Bivariate Scales in R
library(patchwork)            # Composing gpglot2 plots

sessioninfo::session_info()$packages |> 
  as_tibble() |> 
  select(package, 
         version = loadedversion, 
         date, source) |> 
  arrange(package) |> 
  janitor::clean_names(
    case = "title"
  ) |> 
  gt::gt() |> 
  gt::opt_interactive(
    use_search = TRUE
  ) |> 
  gtExtras::gt_theme_espn()

Table 1: R Packages and their versions used in the creation of this page and graphics

References

Wickham, Hadley, Thomas Lin Pedersen, and Dana Seidel. 2023. “Scales: Scale Functions for Visualization.” https://CRAN.R-project.org/package=scales.

Xiao, Nan. 2024. “Ggsci: Scientific Journal and Sci-Fi Themed Color Palettes for ’Ggplot2’.” https://CRAN.R-project.org/package=ggsci.