Flavor physics still shows two sets of discrepancies between data and standard model predictions. The first one involves a transition from b to s and the other one appears in b to c decays. Many models have been build to explain one or both sets of anomalies. One of the main challenges for them is the necessity to produce large contributions to a quark-lepton transition, while avoiding data excluded contributions to quark-quark and lepton-lepton interactions. I will first review these sets of data and briefly present a leptoquark based approach. I will then present a new viable model for b to s transition relying on a gauge horizontal symmetry.