ggplot2：条形图中的 y 轴缩放问题

Question

I wanne create a bar chart using the following code:

ggplot(data_set, aes(x=reorder(regionname, +gdpcap), y=gdpcap)) +
  geom_bar(stat="identity")

As seen in the code, the y-axis doesn't simply display the count but the mean of the variable 'gdpcap' for each category on the x-axis.

In the dataset, the values for the variable 'gdpcap' are continuous and only range from 1 to 10. But in the graphic output of my code, the values on the y-axis are multiplied by ten and display '0', '20', '40', '60' instead of just '0, 2, 4, 6'. It can be seen in the attached picture: bar chart

Why is ggplot scaling my y-axis differently? If I would just calculate the mean, using the following code...

mean(data_set$gdpcap)

...then the output would be 3.681204 and not 30.681204. So something in the ggplot command is causing the issue I assume. Any ideas?

Thanks for any help!

Thomas

Answer 1

As seen in the code, the y-axis doesn't simply display the count but the mean of the variable 'gdpcap' for each category on the x-axis.

No, that's not the case. Nothing in your code tells ggplot to take a mean. stat = "identity" means the y-values are taken as-is. The default position = "stack" means little bars for each observation are stacked on top of each other, so what you're seeing is the sum.

There's not a built-in way to plot means using geom_bar , but we can use stat_summary to have ggplot do the calculation. If it were me though, I'd use dplyr for the data transformation (that's what it's for) and ggplot for the plotting (that's what it's for).

Here it is both ways, demonstrated on the built-in mtcars data:

ggplot(mtcars, aes(x = factor(cyl), y = mpg)) +
  stat_summary(geom = "col", fun = mean)


mtcars %>%
  group_by(cyl) %>%
  summarize(mean_mpg = mean(mpg)) %>%
  ggplot(aes(x = factor(cyl), y = mpg)) +
  geom_col()

(I've switched to geom_col , which is equivalent to geom_bar(stat = "identity") .)