train_session_join

EDA

# Join train and session to find  no of users with log information
left=train.drop_duplicates(subset='id', keep='first')
right = session.drop_duplicates(subset='user_id', keep='first')
merge_id=left.merge(right, left_on='id', right_on='user_id', how='inner')

# print info
print("Train  :",left.shape)
print("Session:",right.shape)
print("Merge  :",merge_id.shape)
print("No of users in Train Data with Session Information:",merge_id.shape[0])
print("{} / {} = {}".format(merge_id.shape[0],left.shape[0],np.round((merge_id.shape[0]/left.shape[0]),2)))