制作用于分布连续变量的条形图

Question

I'm making a barplot to visualize distribution of continuous variable, eg price distribution for listing price. I generated some sample data and make a barplot with ggplot2.

a<- rnorm (100, 1000, 1000)
d <- as.data.frame(a)
d <-d %>%  mutate(b=cut(a, breaks=seq(min(a),max(a), 500))) %>% 
    group_by(b) %>% summarize(count=n())
ggplot(data=d, aes(x=b, y=count)) + 
    geom_bar(stat = 'identity')+
    theme (axis.text.x=element_text(angle=90, size=5, face='bold'))

My question is

• how can I format the xaxis label so that, for example, 1.22e+03 becomes 1220.

• why the last bin becomes NA?

I know I can just use geom_histogram for this data. But I just want to have some flexibility to cut the continuous variable into bins for some highly skewed data. Any help is very much appreciated. Thanks in advance.

Answer 1

Both issues are about cut() . You should read ?cut

To avoid scientific notations in the classes labels, use the argument dig.lab . In your example, cut(a, breaks=seq(min(a),max(a), 500), dig.lab = 6L) seems to be enough.

NA s appear for two reasons linked to your breaks argument. First, by default, the first break is excluded from cut() , so that the observation where a == min(a) will be NA . To overcome this, use include.lowest = TRUE .

Finally, your highest values will be ignored because seq(min(a), max(a), 500) produces a vector that stops at the last multiple of 500 before max(a) and therefore does not include max(a) . To overcome this, you need to make sure the second argument of seq is the first multiple of 500 after max(a) , like ceiling(max(a) / 500) * 500 .

Therefore, this should work:

d <-d %>%  
  mutate(b=cut(a, breaks=seq(min(a), ceiling(max(a) / 500) * 500, 500), 
               include.lowest = TRUE, 
               dig.lab = 6L)) %>% 
  group_by(b) %>% summarize(count=n())