select gender, university, count(device_id) as user_num, avg(active_days_within_30) as avg_active_day, avg(question_cnt) as avg_question_cnt from user_profile group by gender, university