Looking for Fairness in Recommender Systems