Synthesized Anscombe's quartet

synethsized anscombe.R

library(tidyverse)
library(synthpop)

# make the built-in anscombe dataset tidy
anscombe_tidy = anscombe %>%
  rowid_to_column("pointNumber") %>% 
  # pivot long
  gather(setNumber, value, -pointNumber) %>%
  # get the dimention and set number
  mutate(dimention = substring(setNumber, 1, 1), setNumber = substring(setNumber, 2)) %>%
  # combine each point into a row
  spread(dimention, value) %>% 
  # cleanup
  dplyr::select(-pointNumber) %>%
  mutate(setNumber = paste('Set', setNumber))

# plot the data
plot_original = ggplot(anscombe_tidy) +
  aes(x=x, y=y) +
  geom_point(size = 2, color = "#377EB8", alpha = 0.5) +
  stat_smooth(geom='line', method = "lm", color = "#E41A1C", alpha = 0.75, se = FALSE, fullrange = TRUE) +
  facet_grid(. ~ setNumber) +
  theme_bw() +
  labs(title = "Anscombe's Quartet")
  

# synethesize each set in the data
synthesized_anscombe = anscombe_tidy %>% 
  group_by(setNumber) %>% 
  do(synthpop::syn(., seed = 12345)$syn)

# plot the synthesized data
plot_synthesized = ggplot(synthesized_anscombe) +
  aes(x=x, y=y) +
  geom_point(size = 2, color = "#377EB8", alpha = 0.5) +
  stat_smooth(geom='line', method = "lm", color = "#E41A1C", alpha = 0.75, se = FALSE, fullrange = TRUE) +
  facet_grid(. ~ setNumber) +
  theme_bw() +
  labs(title = "Synthesized Anscombe's Quartet")

# plot both together
cowplot::plot_grid(plot_original, plot_synthesized, ncol = 1)