CODEWITHSUNDEEP
rggplot2
Autumn
Autumn

Reputation: 585

how to make plot scales the same when using ggplot2?

I am plotting regression lines using ggplot2. I also want to plot out the mean points with confidence interval to show if it is significant. For example:

library("ggplot2")
library("gridExtra")
library("epicalc")
options(digits=2)

subset1 <- subset(na.omit(iris), Species == "setosa")
subset2 <- subset(na.omit(iris), Species == "versicolor")
subset3 <- subset(na.omit(iris), Species == "virginica")

meanx <- c(ci(subset1$Sepal.Length)$mean,
               ci(subset2$Sepal.Length)$mean,
               ci(subset3$Sepal.Length)$mean)

meany <- c(ci(subset1$Sepal.Width)$mean,
           ci(subset2$Sepal.Width)$mean,
           ci(subset3$Sepal.Width)$mean)

Species <- factor(c("setosa", "versicolor", "virginica"))
meanmatrix <- as.data.frame(cbind(Species, meanx, meany))

lowerx <- c(ci(subset1$Sepal.Length)$lower95ci,
            ci(subset2$Sepal.Length)$lower95ci,
            ci(subset3$Sepal.Length)$lower95ci)

upperx <- c(ci(subset1$Sepal.Length)$upper95ci,
            ci(subset2$Sepal.Length)$upper95ci,
            ci(subset3$Sepal.Length)$upper95ci)

lowery <- c(ci(subset1$Sepal.Width)$lower95ci,
            ci(subset2$Sepal.Width)$lower95ci,
            ci(subset3$Sepal.Width)$lower95ci)

uppery <- c(ci(subset1$Sepal.Width)$upper95ci,
            ci(subset2$Sepal.Width)$upper95ci,
            ci(subset3$Sepal.Width)$upper95ci)

px <- ggplot(data = meanmatrix, geom = 'blank',
             aes(y = meanx, x = meany,color = factor(Species)))
pbx <- px + 
  geom_point(size = 5) +
  geom_errorbar(aes(ymin=lowerx, ymax=upperx), colour="black", width=.1) +
  scale_color_manual(values = c("#00FFFF", "#FFFF00", "#00FF00")) +
  theme(panel.background = element_rect(fill='white', colour='red'),
        axis.title.x = element_blank(),
        axis.title.y = element_blank(),
        legend.position = "none") +
  coord_flip()


py <- ggplot(data = meanmatrix, geom = 'blank',
             aes(y = meany, x = meany,color = factor(Species)))
pby <- py + 
  geom_point(size = 5) +
  geom_errorbar(aes(ymin=lowery, ymax=uppery), colour="black", width=.1) +
  scale_color_manual(values = c("#00FFFF", "#FFFF00", "#00FF00")) +
  theme(panel.background = element_rect(fill='white', colour='red'),
        axis.title.x = element_blank(),
        axis.title.y = element_blank(),
        legend.position = "none")

p <- ggplot(iris, aes(x = Sepal.Length, y = Sepal.Width,
                    color = factor(Species)))

p0 <- p + 
  scale_color_manual(values = c("#00FFFF", "#FFFF00", "#00FF00")) +
  scale_linetype_manual(breaks = c("0","1"), values = c(1,2), labels = c("male", "female")) +
  geom_smooth(method = "lm",se = FALSE, size = 1.2) +
  theme(panel.background = element_rect(fill='white', colour='red'),
        axis.title.x = element_blank(),
        axis.title.y = element_blank(),
        legend.position = "none")

grid.newpage() 
pushViewport(viewport(layout = grid.layout(nrow=3, ncol=3)))
print(p0,vp = viewport(layout.pos.row = 1:2, layout.pos.col = 2:3))
print(pby,vp = viewport(layout.pos.row = 1:2, layout.pos.col = 1))    
print(pbx,vp = viewport(layout.pos.row = 3, layout.pos.col = 2:3))

The scales of the three plots are different. How can I make them universal so that I can compare them? Thanks.enter image description here

Upvotes: 0

Views: 438

Answers (1)

Paul Hiemstra
Paul Hiemstra

Reputation: 60984

Like Ernest A. commented, you can manually change the scales of the x and y-axis using scale_x_continuous and scale_y_continuous. Just set the breaks argument to the same values.

It could also be just easier to plot the uncertainty bars in the plot itself, or plot the regression lines including a 95% confidence interval.

Upvotes: 1

Related Questions