Following graph is from this article: enter image description here

It plots percent of participants showing improvement of different percent cut-off values (or more). This can be applied to a large number of situations where 2 drugs or interventions are given to 2 groups and some parameter is measured before and after study period.

A similar graph plotted for iris dataset is as follows:

enter image description here

Is there any general name for this type of plot/graph?

Thanks for your insight.

  • $\begingroup$ Although the means of construction of the iris data plot is unclear, both of the images look like empirical survival functions. Compare to this plot in R: library(data.table); X <- as.data.table(iris[order(iris$Petal.Length), ]); X[, step := 1/.N, by=Species]; X[, S := 1 - cumsum(step), by=Species]; library(ggplot2); ggplot(X, aes(Petal.Length, S, color=Species)) + geom_step(size=1) + geom_point(data=X[, .(S=min(S)), keyby=.(Species, Petal.Length)]) $\endgroup$ – whuber Sep 2 '20 at 15:42

I know this as an empirical survival plot for the corresponding variable, grouped by another factor. Note that the values plotted are simply $1 - \hat{F_n}(i)$ where $\hat{F_n}(\cdot)$ is the empirical cumulative distribution function.

  • 1
    $\begingroup$ +1. However, a close look at the last plot shows that's not quite how it is constructed. Instead, only points of the form $(x, S(x))$ with $x\in\{0,1/2,1,\ldots,8\}$ have been plotted. $\endgroup$ – whuber Sep 2 '20 at 15:44
  • 1
    $\begingroup$ @whuber you’re right, I guess it might be more correct to say that the plot shows 1-ecdf(x), sampled at regular points and linearly interpolated. $\endgroup$ – Maximilian Aigner Sep 2 '20 at 15:46
  • 1
    $\begingroup$ Other names and more discussion at stats.stackexchange.com/questions/485328/… $\endgroup$ – Nick Cox Sep 2 '20 at 15:58
  • $\begingroup$ It will be very useful if you can add code (R or Python) to properly draw "Emperical Survival Plot". $\endgroup$ – rnso Sep 2 '20 at 17:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.