按累计天数对 geom_tile plot 的 y 轴进行排序

Question

I have a dataset with 4 columns client , date , sales and scale .

I am trying figure it ou how to to order the y axis ( client ) on a geom_tile plot not based on the default decreasing order of the levels but instead on the cumulative sum of each client for all days.

The code bellow is an example. The client are ordered 5,4,3,2,1 but I need instead ordered based on the sales of all days.

data.frame(client=seq(1:5),date=Sys.Date()-0:05,sales=rnorm(30,300,100)) %>% mutate_if(is.numeric,round,0) %>% mutate(escale=cut(sales,breaks=c(0,100,200,300,1000),labels=c("0-100","100-200","200-300","+300"))) %>% ggplot(.,aes(x=date,y=client,fill=escale)) + geom_tile(colour="white",size=0.25)

Appreciate any help

Answer 1

Maybe a quick fix like this?

df <- data.frame(client=seq(1:5),date=Sys.Date()-0:05,sales=rnorm(30,300,100)) %>% 
mutate_if(is.numeric,round,0) %>% 
mutate(escale=cut(sales,breaks=c(0,100,200,300,1000),
labels=c("0-100","100-200","200-300","+300")))

# we get the order according to total sales
o <- names(sort(tapply(df$sales,df$client,sum)))

ggplot(df,aes(x=date,y=client,fill=escale)) + 
geom_tile(colour="white",size=0.25)+
# just manually set the y-axis here
scale_y_discrete(limits=o)

Personally, I prefer the set the factor levels, in the data.frame, and plot

# we get the order according to total sales
o <- names(sort(tapply(df$sales,df$client,sum)))

df %>% mutate(client = factor(client,levels=o)) %>%
ggplot(.,aes(x=date,y=client,fill=escale)) + 
geom_tile(colour="white",size=0.25)

Both will give this: