Keep order in plot of pandas groupby

Question

I have a dataframe that looks like the below

univ  date            ms      cs
 C    11/01/1977      0.2     0.1
 C    11/02/1977      0.3     0.4
 B    11/01/1977      0.6     -0.1
 B    11/02/1977      0.55    -0.5
 A    11/01/1977      0.23    0.2
 A    11/02/1977      0.12    0.15
 etc

This is what I currently have

fig, axes = plt.subplots(1, 3, figsize=(15, 5))

for (universe, group), ax in zip(df.groupby([df.univ]), axes.flatten()):
    group.plot(x='ms', y='cs', kind='scatter', ax=ax, title=universe)
    ax.set_xlim(0, 1)
    ax.set_ylim(0, 1)

This produces 3 scatter plots, with limits from 0 to 1. What I would like to do is keep the order of the plots titled as C, B, A, such that it gets ordered by A,B,C.

Thanks for your help!

Possible duplicate of [How to overplot a line on a scatter plot in python?](https://stackoverflow.com/questions/19068862/how-to-overplot-a-line-on-a-scatter-plot-in-python) — Andrew L, Jul 26 '17 at 23:16

Alexander · Accepted Answer · 2017-07-26T23:34:26.897

The problem is with the groupby which returns the items in sorted order by default. Just set the flag sort=False.

for (universe, group), ax in zip(df.groupby([df.univ], sort=False), axes.flatten()):
    group.plot(x='ms', y='cs', kind='scatter', ax=ax, title=universe)
    ax.set_xlim(0, 1)
    ax.set_ylim(0, 1)

You could also return them in any order you desired.

gb = df.groupby('univ')
desired_order = ('B', 'A', 'C')
for (universe, group), ax in zip(((u, gb.get_group(u)) for u in desired_order), 
                                 axes.flatten()):
    ...

And btw, you should only have one question per post.

Keep order in plot of pandas groupby

1 Answers1