# You can – optionally – remove the unnecessary columns and # keep the user_id column only: article_read.groupby('source').count()[['user_id']] # Change user_id as you see fit.