Add axis-reversed stat_ecdf #3832
Conversation
* fix tidyverse#3727 * remove error entirely
…ilable on some layers (tidyverse#3757)
Add an option to reverse axis y->x instead of x->y to acomodate tail-latency plots, i.e.: https://github.com/ixy-languages/ixy-languages/blob/master/img/latency-hdr-hist-1.png.
… from data (tidyverse#3789) * improve lims docs: add warning about values outside of limits being removed from data, references tidyverse#3752 * place lim warning in first paragraph, references tidyverse#3752 * rerun roxygen, references tidyverse#3752
* testing * experimenting * update annotation_logticks let ticks be outside of plot * fix "outside" option * formatting text * news file update * news file update * news.md update * reconcile News.md with recent updates * format FALSE
…#3837) * Let check_nondata_cols allow Vector class * Remove as_tibble from as_gg_data_frame * Add NEWS.md bullet * Rephrase news bullet
Add an option to reverse axis y->x instead of x->y to acomodate tail-latency plots, i.e.: https://github.com/ixy-languages/ixy-languages/blob/master/img/latency-hdr-hist-1.png.
|
stat_ecdf now uses the now aes switching scheme, for which I added a test based on the test for stat_density. Furthermore, the branch was rebased on top of master. |
|
@jgjl Just an FYI: There's no need to rebase. Merge is fine. We squash-merge all PRs anyways. |
Should I create a new PR without the rebase? |
|
No, no need to change this now. I just wanted to point out that rebasing is not necessary, since it can be a lot of unnecessary extra work. |
|
@jgjl I think you'll have to create a new PR. This PR now lists 60 commits and 49 files changed and it just looks like things went wrong somewhere along the way. Also, I suggest you open an issue first describing the feature you'd like to add, and then you can make a new PR implementing that feature. |
Add an option to reverse axis y->x instead of x->y to acomodate tail-latency plots, i.e.: https://github.com/ixy-languages/ixy-languages/blob/master/img/latency-hdr-hist-1.png (context: https://github.com/ixy-languages/ixy-languages).
In this kind of plot the probability is plotted along the horizontal axis while the values are plotted along the vertical axis. Hence, in comparison to the default stat_ecdf function, the axis are reversed.
This specific step could also be achived by applying coord_flip. However, in this case the reversed-logarithmic transformation of the x-axis cannot be applied. Hence, in my understanding, switching axis in stat_ecdf is the only way to get there.
I will supply the complementary transformation object for the x-axis separately outside of the ggplot2 package.
The proposed implementation is a little simplistic in that it just adds a version where input and output variables are reversed. An alternative would be to allow the input and output variables to be freely choosen. I am happy to supply such a version if it would be preferrable.