Skip to content

Latest commit

 

History

History
112 lines (79 loc) · 4.59 KB

ggplot-on-fire.md

File metadata and controls

112 lines (79 loc) · 4.59 KB
Raw

Portrait of a ggplot on fire 🔥

Picture taken from Pyramide Films.

Extracting colors from images with the colorfindr package

This is a short tutorial that will get you started with building your very own custom color palette with the {colorfindr} package by David Zumbach, and implement it in a ggplot. For this we will be using the wonderful Palmer Penguins package by Allison Horst, Alison Hill and Kristen Gorman. There are many, many color palettes already out there, but what I love about colorfindr is that it is 100% customizable. You can choose to extract however many colors you want and use any favorite film, TV show or art work as inspiration.

Step 1: Install colorfindr

# install.packages("colorfindr") 
# install.packages("tidyverse") 
# install.packages("palmerpenguins")
library(colorfindr) # extracts colors from images
library(tidyverse) # tidy data wrangling & plotting

Step 2: Extract colors from images

# Get colors and create a palette with n = 5 
get_colors("images/hemadramamaisexterieure4612433621.jpg") %>% # you can reference a local file on your computer or a jpg web address
  make_palette(n = 5) # here we extract 5 colors

## [1] "#7EBDB5" "#338791" "#CCD9BD" "#1D1512" "#714F34"

You can play around with how many colors you extract - this will depend on what your data looks like. Do you want to show a few different categories or continuous values? Tip: For very expressive color palettes, I would pick images with strong, vibrant colors. For example:

If you’ve decided that you want 3 different colors, save your new palette to a so-called character vector. You can also ‘set a seed’, which will ensure that you get the same colors every time you run your script.

# Use your favorite magic number to set a seed:
# set.seed(42)
PLF<- get_colors("images/hemadramamaisexterieure4612433621.jpg") %>% 
  make_palette(n = 3)

Step 3: Use your new palette

From this:

library(palmerpenguins) # get the penguin data
penguins %>% # feed ggplot the data
  ggplot(aes(flipper_length_mm, fill = species)) + # we want to investigate the flipper length of our penguin friends
  geom_density(alpha = 0.8, color =NA)+ # specify how you want to display the data
  theme_minimal() # pick a theme

To this 🔥:

penguins %>%
  ggplot(aes(flipper_length_mm, fill = species)) +
  geom_density(alpha = 0.8, color =NA)+
  scale_fill_manual(values = PLF)+
  theme_minimal()

Get inspired by the cool palettes other folks made with colorfindr

Kelly Cotton built a new package of New York-inspired palettes: https://github.com/kellycotton/nycpalettes

Ben Stanley made a palette for Kieslowski's "Dekalog":https://twitter.com/BDStanley/status/1307605257019629570

Lydia Hopper made a palette inspired by Japanese macaques: https://twitter.com/LydiaMHopper/status/1308403623592943616

Kaylea was inspired by #TidyTuesday and looked at reasons for expedition terminations: https://twitter.com/KayleaHaynes/status/1308799804357709825

John MacKintosh made his own "Metallicart" package with album covers of Metallica as inspiration: https://github.com/johnmackintosh/metallicaRt

Kelly Morrow McCarthy was inspired by an abstract photo: https://twitter.com/KellyMM_neuro/status/1308763829095063560

References.