The LEGO Colour Explosion

TidyTuesday Visualisation by Cedric Scherer

NoteAbout this showcase

This is an adaptation of Cedric Scherer’s award-winning TidyTuesday LEGO visualisation. It uses ggsankey to show how the variety of LEGO brick colours exploded over the last two decades. This page requires R with several specialist packages — it is pre-rendered locally via freeze: auto.

library(tidyverse)
library(ggsankey)
library(ggtext)
library(colorspace)
library(patchwork)
library(systemfonts)

col_dark   <- "#353548" #3d3d53, 373759
col_light  <- "#f3efeb"
col_dark2  <- darken(col_dark, .2)
col_light2 <- darken(col_light, .15)

theme_set(theme_void(base_size = 18, base_family = "Pally"))
theme_update(
  axis.text.x = element_text(
    color = col_dark2, face = "bold", margin = margin(15, 0, 15, 0)
  ),
  panel.grid.major.x = element_line(
    color = col_dark2, linetype = "13", size = .9
  ),
  legend.position = "none",
  panel.background = element_rect(fill = col_dark, color = col_dark),
  plot.background = element_rect(fill = col_light, color = col_light),
  plot.margin = margin(0, 0, 0, 0)
)

Data

df_colors <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2022/2022-09-06/colors.csv.gz')
df_inventory_parts <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2022/2022-09-06/inventory_parts.csv.gz')
df_inventories <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2022/2022-09-06/inventories.csv.gz')
df_sets <-  readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2022/2022-09-06/sets.csv.gz')

## code snippet partly by Cara Thompson
## https://github.com/cararthompson/tidytuesdays/blob/master/scripts/202209_lego.R
df_lego <- 
  df_inventory_parts %>% 
  left_join(df_colors, by = c("color_id" = "id")) %>%
  left_join(df_inventories, by = c("inventory_id" = "id")) %>%
  left_join(
    df_sets %>% 
      rename(set_name = name) %>%
      filter(!stringr::str_detect(set_name, "Modulex|modulex")) %>% 
      select(set_num, set_name, year, theme_id), 
    by = "set_num"
  ) %>%
  left_join(
    tibble(
      name = c(df_colors %>% pull("name")),
      basic_name = c(df_colors %>% pull("name") %>% 
        stringr::word(., -1) %>% 
        gsub("Trans-|Black-|DBGray|-", "", .))
    )
  ) %>%
  mutate(hex = paste0("#", rgb)) %>%
  filter(!is.na(year)) %>% 
  group_by(hex, year) %>% 
  summarize(quantity = sum(quantity)) %>% 
  ungroup() %>% 
  complete(year = full_seq(year, 1), hex, fill = list(quantity = 0)) %>% 
  group_by(hex) %>% 
  mutate(n = sum(quantity)) %>% 
  ungroup()

threshold <- 100000

df_lego_main <- 
  df_lego %>%
  mutate(group = if_else(n < threshold, "#13e727", hex)) %>% 
  #filter(year > 1958, year < 2022) %>%
  filter(year < 2022) %>% 
  group_by(group, year) %>% 
  summarize(quantity = sum(quantity))

df_lego_rare <- 
  df_lego %>%
  filter(n < threshold) %>% 
  filter(year > 1977 & year < 2022)

Plot

annotation <- tibble(
  x = 1952, y = 135000, 
  label = "Bricks with <b style='color:#13e727;'>rare colors</b> became much more common over the last two decades."
)

p_main <-
  df_lego_main %>% 
  ggplot(aes(x = year, node = group, value = quantity, fill = group)) +
  annotate(
    geom = "rect", xmin = -Inf, xmax = Inf, ymin = 12000, ymax = Inf,
    fill = col_dark, color = col_dark
  ) +
  geom_sankey_bump(color = "transparent", space = 2000, smooth = 8) +
  geom_sankey_bump(color = "transparent", space = 2000, smooth = 8, alpha = .5) +
  geom_textbox(
    data = annotation, aes(x = x, y = y, label = label), 
    color = NA, fill = NA, text.colour = col_light2,
    family = "Pally", size = 5.3, fontface = "plain", lineheight = .95,
    width = unit(4.3, "inch"), hjust = 0, halign = 0, 
    inherit.aes = FALSE, stat = "unique"
  ) +
  scale_x_continuous(expand = c(0, 0), breaks = seq(1950, 2020, by = 10)) +
  scale_fill_identity() + 
  theme(
    axis.text.x = element_text(color = col_light2, size = 9, margin = margin(8, 0, 8, 0)),
    panel.grid.major.x = element_line(size = .6),
    plot.background = element_rect(fill = col_dark2, color = col_dark2),
    plot.margin = margin(10, 10, 0, 10)
  )

title <- tibble(
  x = 1979, y = 57000, 
  label = paste0('<b style="color:', col_light, ';font-size:62pt;">The LEGO Color Explosion</b><br><br>While in the early days LEGO bricks were limited to a few main colors, the variety of unique colors increased remarkably over the last two decades as the company started releasing a larger number of sets and themes.<br>The main visualization shows the number of parts with "rare colors" across all LEGO sets over time, ranked by the share of parts featured in all newly released sets. The inset plot shows the share of these rare colors in comparison to the most common main colors.<br><br><b style="font-size:11pt;">Graphic: Cédric Scherer • Data: rebrickable</b>')
)

p_rare <- 
  df_lego_rare %>% 
  ggplot(aes(x = year, node = hex, value = quantity, fill = hex)) +
  annotate(
    geom = "rect", xmin = -Inf, xmax = Inf, ymin = 11500, ymax = Inf,
    fill = col_dark, color = col_dark
  ) +
  geom_sankey_bump(color = "transparent") +
  geom_sankey_bump(color = "transparent", alpha = .5) +
  geom_textbox(
    data = title, aes(x = x, y = y, label = label), 
    color = NA, fill = NA, text.colour = col_light2,
    family = "Pally", size = 5.3, lineheight = 1.25,
    width = unit(9.8, "inch"), hjust = 0, halign = 0, vjust = 1, 
    inherit.aes = FALSE, stat = "unique"
  ) +
  scale_x_continuous(expand = c(0, 0), breaks = seq(1980, 2020, by = 5)) +
  scale_y_continuous(expand = c(0, 0), limits = c(NA, 64000)) +
  scale_fill_identity()

panel <- p_rare + inset_element(p_main, left = .03, bottom = .06, right = .41, top = .41)

panel

TipOriginal visualisation

The original by Cedric Scherer uses custom fonts (Pally) and higher resolution rendering. See the original code on GitHub for the full production version.