Without scale_y_log10, geom_signif works as expected, e.g.
ggplot(mpg, aes(class, hwy)) + geom_boxplot() +
geom_signif(annotations = c("First", "Second"), y_position = c(30, 40), xmin = c(4, 1), xmax = c(5, 3))
With scale_y_log10, y-axis become extremely large and original plots are squeezed to the bottom
ggplot(mpg, aes(class, hwy)) + geom_boxplot() +
geom_signif(annotations = c("First", "Second"), y_position = c(30, 40), xmin = c(4, 1), xmax = c(5, 3)) + scale_y_log10()
geom_signif affects scale_y_log10's behavior:
Without scale_y_log10, geom_signif works as expected, e.g. ggplot(mpg, aes(class, hwy)) + geom_boxplot() + geom_signif(annotations = c("First", "Second"), y_position = c(30, 40), xmin = c(4, 1), xmax = c(5, 3))
With scale_y_log10, y-axis become extremely large and original plots are squeezed to the bottom ggplot(mpg, aes(class, hwy)) + geom_boxplot() + geom_signif(annotations = c("First", "Second"), y_position = c(30, 40), xmin = c(4, 1), xmax = c(5, 3)) + scale_y_log10()