R/ggplot Cumulative Sum in Histogram

Question

I have a dataset with user IDs and the number of objects they created. I drew the histogram using ggplot and now I'm trying to include the cumulative sum of the x-values as a line. The aim is to see much the bins contribute to the total number. I tried the following:

ggplot(data=userStats,aes(x=Num_Tours)) + geom_histogram(binwidth = 0.2)+
   scale_x_log10(name = 'Number of planned tours',breaks=c(1,5,10,50,100,200))+
   geom_line(aes(x=Num_Tours, y=cumsum(Num_Tours)/sum(Num_Tours)*3500),color="red")+
   scale_y_continuous(name = 'Number of users', sec.axis = sec_axis(~./3500, name = "Cummulative percentage of routes [%]"))

This does not work because I don't include any bins so the plot

and

ggplot(data=userStats,aes(x=Num_Tours)) + geom_histogram(binwidth = 0.2)+
   scale_x_log10(name = 'Number of planned tours',breaks=c(1,5,10,50,100,200))+
   stat_bin(aes(y=cumsum(..count..)),binwidth = 0.2, geom="line",color="red")+
   scale_y_continuous(name = 'Number of users', sec.axis = sec_axis(~./3500, name = "Cummulative percentage of routes [%]"))

Resulting in this: Result 1 .

Here the cumsum of the count is considered. What I want is the cumsum of the count * value of the bin. Then it should be normalized, so that it can be displayed in one plot. What I am trying to to is something like that:

Example

I would appreciate any input! Thanks

Edit: As test data, this should work:

userID <- c(1:100)
Num_Tours <- sample(1:100,100)
userStats <- data.frame(userID,Num_Tours)
userStats$cumulative <- cumsum(userStats$Num_Tours/sum(userStats$Num_Tours))

example data please

mtoto
– mtoto

2017-06-04 12:16:31 +00:00
Commented Jun 4, 2017 at 12:16 — mtoto
– mtoto, Commented Jun 4, 2017 at 12:16

Marco Sandri · Accepted Answer · 2017-06-04 13:56:36Z

2

Here is an illustrative example that could be helpful for you.

set.seed(111)
userID <- c(1:100)
Num_Tours <- sample(1:100, 100, replace=T)
userStats <- data.frame(userID, Num_Tours)

# Sorting x data
userStats$Num_Tours <- sort(userStats$Num_Tours)
userStats$cumulative <- cumsum(userStats$Num_Tours/sum(userStats$Num_Tours))

library(ggplot2)
# Fix manually the maximum value of y-axis
ymax <- 40
ggplot(data=userStats,aes(x=Num_Tours)) + 
   geom_histogram(binwidth = 0.2, col="white")+
   scale_x_log10(name = 'Number of planned tours',breaks=c(1,5,10,50,100,200))+
   geom_line(aes(x=Num_Tours,y=cumulative*ymax), col="red", lwd=1)+
   scale_y_continuous(name = 'Number of users', sec.axis = sec_axis(~./ymax, 
    name = "Cumulative percentage of routes [%]"))

answered Jun 4, 2017 at 13:56

Marco Sandri

24.5k7 gold badges60 silver badges64 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

R/ggplot Cumulative Sum in Histogram

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related