Change order of columns in stacked bar plot

Question

I'm using the following code to produce a stacked bar plot from a pandas data frame in Python ( data is the DataFrame object).

data.plot(kind='bar', stacked=True, figsize=(15, 10), x='Species', fontsize=16,
          rot=0)

Here is an IPython Notebook for context.

I would like to change the order of the columns. The seaborn.barplot function has the x_order argument, which does exactly what I would like. However, it does not support stacked bar plots directly like the pandas.DataFrame.plot function does. What is the simplest way to reorder the columns?

Answer 1

You can manually rearrange the rows by permuting the numeric indices, like the answer given by @iayork, or alternatively you can set the index of your dataframe to be the column Species .

species = ['A', 'B', 'H', 'S', 'P', 'N']

data2 = data.set_index(['Species'])
data2.reindex(species).plot(kind='bar', stacked=True, figsize=(15, 10),
                            fontsize=16, rot=0)

By setting the index of data as Species , you can now permute the rows by their string labels, instead of numeric indices.

Answer 2

The simplest way is probably to rearrange your index.

data2 = data.loc[[5, 4, 3, 2, 1, 0]]

reverses the order in dataframe and plot.