I have a dataset in pandas, with four columns (year, month, day, register).

df_registration

data table image

I want to group the data by year and month, and then count for each month in a year how many 'yes' and how many 'no' there were (or at least how many 'yes').

Image of df with expected result:

desired outcome

I have tried doing group_by count,

g = df_registration.groupby(["year", "month"])
monthly_counts = g.aggregate({"register": pd.Series.value_counts })

but the outputs do not give the desired outcome, they just count the number of both register values.

Image of df with failed attempt:

wrong

I cannot get it to work as I want...

EDIT!____________ solution _________________

The code from alex smolyakov in the comment works!

counts = df_registration.groupby(["year", "month"]["register"].value_counts()

the code output here

CodePudding user response：

counts = df_registration.groupby(["year", "month"])["register"].value_counts()

does it help ?