Pandas plot axvspan

Question

I have a dataframe like this

from datetime import date
import random

start_date = date.today().replace(day=1, month=1).toordinal()
end_date = date.today().toordinal()

df=pd.DataFrame.from_dict({'dates':[ date.fromordinal(random.randint(start_date, end_date)) for i in xrange(10) ]
                        ,'n':np.random.randint(0,100,10)})
df.index=df['dates']; del df['dates']
df.sort_index(inplace=True)

Which generates

             n
dates         
2016-01-24   0
2016-02-26  98
2016-03-04  68
2016-03-26  45
2016-05-03  89
2016-05-09  83
2016-05-11  58
2016-07-29  53
2016-09-18  79
2016-10-20  57

I would like to plot it and have a axvspan for February and March. Since my dates are random how can I select those 2 months and use axvspan

my plotting code is simple

ax=df.plot()

Does axvspan(pd.datetime(2016, 2, 1), pd.datetime(2016,4,1)) give you what you want? — Brian Huey, Nov 28 '16 at 20:53
That works. For some reason when I do ax.axvspan(pd.datetime(2016, 2, 1), pd.datetime(2016,4,1)) inside a loop I don't see anything. Otherwise it works — NinjaGaiden, Nov 28 '16 at 20:59

score 3 · Answer 1 · edited May 23 '17 at 12:13

Here is an example. It has to be improved if the dates span over several years. You can also check the answer to this question since the principle is exactly the same.

ax = df.plot()
# Getting the boundaries for dates in feb and march
period = df[(df.index.month > 1) & (df.index.month < 3)].index
# Highlighting
ax.axvspan(min(period) - MonthBegin(), max(period) + MonthEnd(),
           facecolor='g', edgecolor='none', alpha=.2)

Pandas plot axvspan

1 Answers1