CODEWITHSUNDEEP
pythonpandasmatplotlib
LondonRob
LondonRob

Reputation: 78723

Retaining category order when charting/plotting ordered categorical Series

Consider the following categorical pd.Series:

In [37]: df = pd.Series(pd.Categorical(['good', 'good', 'ok', 'bad', 'bad', 'awful', 'awful'], categories=['bad', 'ok', 'good', 'awful'], ordered=True))

In [38]: df
Out[38]:
0     good
1     good
2       ok
3      bad
4      bad
5    awful
6    awful
dtype: category
Categories (4, object): [bad < ok < good < awful]

I want to draw a pie or bar chart of these. According to this SO answer, plotting categorical Series requires me to plot the value_counts, but this does not retain the categorical ordering:

df.value_counts().plot.bar()

How can I plot an ordered categorical variable and retain the ordering? Bar chart

Upvotes: 2

Views: 330

Answers (1)

jezrael
jezrael

Reputation: 862611

You can add parameter sort=False in value_counts:

print df.value_counts()
awful    2
good     2
bad      2
ok       1
dtype: int64

print df.value_counts(sort=False)
bad      2
ok       1
good     2
awful    2
dtype: int64

print df.value_counts(sort=False).plot.bar()

graph

Or add sort_index:

print df.value_counts().sort_index()
bad      2
ok       1
good     2
awful    2
dtype: int64

print df.value_counts().sort_index().plot.bar()

Upvotes: 3

Related Questions