You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The default colour bar width strikes me as too thin, in that it can be fairly to very difficult to see the colours inside it depending on how large you are viewing the plot ('very' when it is smaller) and therefore not particularly readable or clear. Morevoer, other visualisation tools I have seen in use in the field have thicker colour bars, in some cases much thicker such as Iris (see https://scitools-iris.readthedocs.io/en/stable/generated/gallery/index.html for the gallery), which suggests it is normal and preferred that way.
I asked a sample of colleagues whether they agreed the default width is too thin and they overwhelmingly (i.e. all surveyed) did. I included plots varying the width to try to gauge what the new default should be, and it seems like colorbar_thick should be increased to 0.035 from the current default of 0.015.
New default width judegement
I did a quick survey via NCAS Slack with a representative (two-line snippet from the docs example 1) with colorbar_thick varying from the default 0.015 then 0.02 to 0.04 in 0.01 increments (see below), and about half thought the 0.03 plot had the most suitable width, and about half the 0.04 width (including people who thought either of those two were best), hence I think inbetween is the best choice, 0.035.
Though I show a contour plot to illustrate, the width default increase will need to be implemented across all plots for universal benefit of the increase and for consistency.
Code and outputs to compare
I varied the colorbar_thick keyword from the default as follows:
Oh, well thanks for registering your thoughts - maybe I need to do a more comprehensive survey, since I am going by the opinions of ~10 people? (BTW 0.015 is the default so it sounds like you still prefer the width to increase, just not by so much?)
The default colour bar width strikes me as too thin, in that it can be fairly to very difficult to see the colours inside it depending on how large you are viewing the plot ('very' when it is smaller) and therefore not particularly readable or clear. Morevoer, other visualisation tools I have seen in use in the field have thicker colour bars, in some cases much thicker such as Iris (see https://scitools-iris.readthedocs.io/en/stable/generated/gallery/index.html for the gallery), which suggests it is normal and preferred that way.
I asked a sample of colleagues whether they agreed the default width is too thin and they overwhelmingly (i.e. all surveyed) did. I included plots varying the width to try to gauge what the new default should be, and it seems like
colorbar_thick
should be increased to0.035
from the current default of0.015
.New default width judegement
I did a quick survey via NCAS Slack with a representative (two-line snippet from the docs example 1) with
colorbar_thick
varying from the default0.015
then0.02
to0.04
in0.01
increments (see below), and about half thought the0.03
plot had the most suitable width, and about half the0.04
width (including people who thought either of those two were best), hence I think inbetween is the best choice,0.035
.Though I show a contour plot to illustrate, the width default increase will need to be implemented across all plots for universal benefit of the increase and for consistency.
Code and outputs to compare
I varied the
colorbar_thick
keyword from the default as follows:where the outputs are:
0.015
, current default0.02
0.03
0.04
The text was updated successfully, but these errors were encountered: