Setting the order of boxes in sns.boxplot() about seaborn HOT 7 CLOSED

Uh oh, that sounds like a bug. Let me look into it.

from seaborn.

If you need sample data to replicate, please let me know. I am happy to
provide an IPython notebook and the data file, but I would rather not
put it up on Github just yet. :)

On 21 Nov 2013, at 12:08, Michael Waskom wrote:

Uh oh, that sounds like a bug. Let me look into it.

---

Reply to this email directly or view it on GitHub:
#14 (comment)

from seaborn.

Indeed, it looks like the boxes end up in the alphabetical order of the grouping variable, but the xlabel names end up in the order they appear in the Series. I have replicated it in this gist.

For the time being, you can specify the names as a parameter (demonstrated in the final cell of the example notebook).

Thanks for reporting this! I'll open a PR to fix it tonight.

from seaborn.

You are fast, my friend. Like lightning.

On 21 Nov 2013, at 12:19, Michael Waskom wrote:

Indeed, it looks like the boxes end up in the alphabetical order of
the grouping variable, but the xlabel names end up in the order they
appear in the Series. I have replicated it in this
gist.

For the time being, you can specify the names as a parameter
(demonstrated in the final cell of the example notebook).

Thanks for reporting this! I'll open a PR to fix it tonight.

---

Reply to this email directly or view it on GitHub:
#14 (comment)

from seaborn.

Fixed in master!

from seaborn.

Shall I put in another feature request for the ability to easily specify the order of the boxplot?

from seaborn.

Sure, go ahead and open a separate issue for that. This fix reminded me that there is a lot of gross repeated code in the setup for boxplot and violin that should be abstracted out. I probably will add the ordering when I get to that, otherwise it will just add more messiness.

In the meantime, you could do (extending the gist example):

cols = ["b", "c", "a"]
data = df.x.groupby(df.g).values[cols]
sns.boxplot(data, names=cols)

as a workaround.

from seaborn.