CODEWITHSUNDEEP
pythonpandas
user1872329
user1872329

Reputation: 361

Create a bar plot in pandas with dates on x-axis, one bar for each value in other column

I have the following pandas dataframe:

>>> df
>>> StartDate         Port  Count
2011-08-10 11:07:10   3128  10
2011-08-10 11:07:40     80   1
2011-08-10 11:07:40    443   1
2011-08-10 11:07:40   3128  10
2011-08-10 11:08:00    443   1
2011-08-10 11:08:00   3128   9
2011-08-10 11:08:20     80   1

I want to create a histogram where on x-axis will be 'StartDate' periods, on y-axis will be count and there will be one column for each value from the 'Port' column.

I tried using groupby() with df.plot.bar(), but it does not give me the result I want. How could I do it?

Ok, not the best drawing, but should give the idea. y-axis is a count, each bar represents the value from the 'Port' column Ok, not the best drawing, but should give the idea. y-axis is a count, each bar represents the value from the 'Port' column. On x-axis are dates from the first column

Upvotes: 2

Views: 540

Answers (1)

jezrael
jezrael

Reputation: 862431

I believe you need set_index + unstack for reshaping and last use DataFrame.plot.bar:

df1 = df.set_index(['StartDate','Port'])['Count'].unstack(fill_value=0)
print (df1)
Port                 80    443   3128
StartDate                            
2011-08-10 11:07:10     0     0    10
2011-08-10 11:07:40     1     1    10
2011-08-10 11:08:00     0     1     9
2011-08-10 11:08:20     1     0     0

df1.plot.bar()

Alternative solution with pivot:

df1 = df.pivot(index='StartDate', columns='Port', values='Count')
print (df1)
Port                 80    443   3128
StartDate                            
2011-08-10 11:07:10   NaN   NaN  10.0
2011-08-10 11:07:40   1.0   1.0  10.0
2011-08-10 11:08:00   NaN   1.0   9.0
2011-08-10 11:08:20   1.0   NaN   NaN

df1.plot.bar()

Upvotes: 1

Related Questions