CODEWITHSUNDEEP
rggplot2
Lorenzo
Lorenzo

Reputation: 11

How can I build cross error bar with ggplot2?

I'm looking a way to build in ggplot a graph like this: enter image description here

I've build horizontal bars with percentage, but i don't know how to make error bars crossing Now my plot is like this: enter image description here

Thanks everyone

Upvotes: 0

Views: 90

Answers (1)

Allan Cameron
Allan Cameron

Reputation: 174278

You didn't provide a reproducible example, so I have no idea how your data is laid out. I have had to guess to try to recreate your data and your plot:

library(ggplot2)

df <- data.frame(Domanda = rep(c("Utile", "Totale", "Intuitiva", "Interessante",
                                 "Chiara", "Bella"), each = 2),
                 voto = factor(rep(c("[1-3]", "[4-6]"), 6),  
                               levels = c("[4-6]", "[1-3]")),
                 `Popolazioni_relativa` = c(13, 87, 15, 85, 25, 75, 8,
                                            92, 30, 70, 30, 70)/100,
                 CI_95_lower = c(4,  4, 6,  6, 14, 14, 2, 2, 16, 16, 16, 16)/100,
                 CI_95_upper = c(20, 20, 25, 25, 35, 35, 16, 16, 40, 40, 40, 40)/100)

ggplot(df, aes(Domanda, `Popolazioni_relativa`, fill = voto)) + 
  geom_col() + 
  coord_flip() + 
  geom_errorbar(aes(ymin = CI_95_lower, ymax = CI_95_lower), position = "identity") +
  geom_errorbar(aes(ymin = CI_95_upper, ymax = CI_95_upper), position = "identity") +
  labs(y = "Popolazio relativa")

There is no geom in ggplot2 to do what you want, so you have to do it yourself with polygons:

errordf <- data.frame(x = rev(do.call(c, lapply(0:5, `+`, c(.56, .8, 1, 1, 1.2, 1.45)))),
                      y = c(13, 4, 13, 13, 20, 13,
                            15, 6, 15, 15, 25, 15,
                            25, 14, 25, 25, 35, 25,
                            8, 2, 8, 8, 16, 8,
                            30, 16, 30, 30, 40, 30,
                            30, 16, 30, 30, 40, 30)/100,
                      voto = factor(rep(rep(c("[4-6]", "[1-3]"), each = 3), 6),
                                    levels = c("[4-6]", "[1-3]")),
                      groups = rep(1:12, each = 3))


ggplot(df, aes(Domanda, `Popolazioni_relativa`, fill = voto)) + 
  geom_col() + 
  coord_flip() + 
  labs(y = "Popolazio relativa") +
  geom_polygon(data = errordf, aes(x, y, group = groups), position = "identity") +
  scale_fill_manual(values = c("#D03020", "#3060C0"))

Created on 2020-06-13 by the reprex package (v0.3.0)

Upvotes: 1

Related Questions