Plot controls about dabest-python HOT 5 CLOSED

Also, is there a way to edit the x-axis tick labels? My variable names aren't pretty and I don't like spaces in them so they don't read nicely.

Thanks

from dabest-python.

Is there a recommended way of saving high-res plots with axis titles that are readable.

The easiest way to do this is with seaborn's set function.

import seaborn as sns
sns.set(style='ticks', font_scale=1.5)

This, however, will scale the font for all text features. To specifically change the axis titles, directly assign the desired font size to matplotlib's rcParams.

import matplotlib.pyplot as plt
plt.rcParams['axes.labelsize'] = 16

Note that this will set global options for any figures created; if you are generating figures that need a different label size, you will need to reset/reassign the label size before plotting.

Also, is there a way to edit the x-axis tick labels? My variable names aren't pretty and I don't like spaces in them so they don't read nicely.

Consider re-naming your group names in pandas before plotting. To directly 'edit' these, you will need to access the appropriate Axes (see here) and use matplotlib (see here for more details) :

f = two_groups_unpaired.mean_diff.plot();

rawdata_axes = f.axes[0]
effsize_axes = f.axes[1]

rawdata_axes.set_xticklabels(["blue group", "orange!"])
effsize_axes.set_xticklabels(["", "my difference"]) # The xtick is actually plotted at position "1".

Thanks,
Joses

from dabest-python.

Excellent. Thanks @josesho .

from dabest-python.

Sorry, last question on this @josesho ( I know it's closed but oh well)... What about removing the legend. I'm using colors for contrast but not to differentiate things to where a legend is needed. Is there anyway to get rid of it. For reference, I'm trying to do the multi-axis plot in the tutorial and the legends end up underneath the other plots.

Thanks

from dabest-python.

Nevermind... took too long but figured it out.

topleft_axes = axx.flat[0]
topleft_axes.contrast_axes.get_legend().remove()

from dabest-python.