python - How to rearrange the columns of a DataFrame into an indexed binary matrix?

Question

I am trying to convert this dataframe:

data_in = {
    'dates': [2017, 2017, 2018, 2019, 2019, 2019],
    'names': ['Roger', 'Rafa', 'Roger', 'Rafa', 'Novak', 'Dom']
}
df_in = pd.DataFrame(data_in)

>>> df_in
    dates   names
0   2017    Roger
1   2017    Rafa
2   2018    Roger
3   2019    Rafa
4   2019    Novak
5   2019    Dom

into this binary matrix:

>>> df_out

    Roger   Rafa    Novak   Dom
dates               
2017    1   1   0   0
2018    1   0   0   0
2019    0   1   1   1

with all the dates as index, all the names as columns, and the data being 1 if occurence of the name at the date, and np.NaN or 0 if not.

I can build the df_out dataframe with its index and columns, but how would you get the data?

Answer 1

Use get_dummies with max for 1 if exist values else 0:

df = pd.get_dummies(df_in.set_index('dates')['names']).groupby(level=0).max()
print (df)
       Dom  Novak  Rafa  Roger
dates                         
2017     0      0     1      1
2018     0      0     0      1
2019     1      1     1      0