How to add or eliminate scatter points on a graph mechanically?

Question

Hello, good morning!

I would like to know a routine to add or eliminate scatter points on a graph in a "mechanical" way with the mouse click on the graph!

I intend to work with some maps, to mark some specific points, but visually on the graph. There is a long time ago I saw somewhere a script that allowed to do this with respect to eliminate the discrepant points in a graph of normal distribution QQ, just clicking on the points on the graph and these were removed... but I can’t remember where else I saw.

Answer 1

Maybe the gatepoints package will be useful to you.

library(gatepoints)
set.seed(1)

df <- data.frame(x = c(rnorm(n = 100, mean = 4, sd = 2), 
                       rnorm(n = 10, mean = 10, sd = 10),
                       rnorm(n = 5, mean = 4, sd = 20)),
                 y = c(rnorm(n = 100, mean = 6, sd = 3), 
                       rnorm(n = 10, mean = 1, sd = 1),
                       rnorm(n = 5, mean = 6, sd = 25)))
{
  grDevices::X11()
  plot(df, col = "red", pch = 16)
  pontos_selecionados <- gatepoints::fhs(df)
}

Select the points of interest.

Click on "stop" in the window.

---

id <- data.frame(id = as.numeric(as.character(pontos_selecionados)),
                 remove = 1)

df2 <- df %>% 
  dplyr::mutate(id = 1:nrow(df)) %>% 
  dplyr::left_join(id, by = "id") %>% 
  replace(is.na(.), 0) %>% 
  dplyr::filter(remove == 0)

ggplot2::ggplot(df2) +
  geom_point(aes(x = x, y = y)) + 
  theme_classic()

df3 <- df %>% 
  dplyr::mutate(id = 1:nrow(df)) %>% 
  dplyr::left_join(id, by = "id") %>% 
  replace(is.na(.), 0) %>% 
  dplyr::mutate(remove = as.factor(remove))


ggplot2::ggplot(df3) +
  geom_point(aes(x = x, y = y,
                 shape = remove, fill = remove), 
             show.legend = F, size = 3, alpha = 0.5) + 
  scale_shape_manual(values = c(21,22)) +
  theme_classic()