2

My data:

df <- data.frame(sp = c(LETTERS[1:8]),
                 tr = c("NS", "LS", "NS", "LS", "LS", "HS", "HS", "HS"),
                 bv = c(14, 5, 11, 5.6, 21, 5.4, 2, 4.8),
                 av = c(0.0, 14, 21, 48.4, 15, 55.6, 37, 66.2))

I do the bar plot

library(reshape2)
df1 <- melt(df, id.vars = c("sp", "tr"))

ggplot(aes(sp, value, fill = variable) , data = df1) + theme_classic() + 
  geom_bar(aes(lty = tr), lwd = 1.2, data = df1, stat = "identity", colour = "black", width =.8) + 
  theme(legend.position = "bottom" ) +
  scale_linetype_discrete(name = "ja") 

Output enter image description here What I does not like is the legend. I'd like to have just the lines type from the second part "ja" and remove "variable" part. I'd like to have the white background in the legend boxes, not the grey one.

Mateusz1981
  • 1,817
  • 17
  • 33

1 Answers1

5

You can remove the variable legend by setting fill = FALSE in guides and you change the backgroundcolor with override.aes in guide_legend (also inside guides) as follows:

ggplot(df1, aes(sp, value, fill = variable)) + 
  geom_bar(aes(lty = tr), lwd = 1.2, stat = "identity", colour = "black", width =.8) + 
  scale_linetype_discrete(name = "ja") +
  guides(fill = FALSE,
         lty = guide_legend(override.aes = list(lty = c('dotted', 'dashed', 'solid'),
                                                fill = "white"))) +
  theme_classic() +
  theme(legend.position = "bottom")

this results in the following plot:

enter image description here

Jaap
  • 81,064
  • 34
  • 182
  • 193