In [7]:

df
Out[7]:
event   event_time  num session_id
0   dog     1   2   a
1   cat     2   3   a
2   bird    3   5   a
3   tree    4   7   a
4   cat     1   3   b
5   dog     2   2   b
6   tree    1   7   c
7   dog     2   2   c
8   cat     3   3   c

pv = pd.pivot_table(df, 'num', rows='event', cols='event', aggfunc=np.sum)