#To be done on a tained model #model here is a stand in for your named model pd.DataFrame( {"feature": X.columns, "importance": model.feature_importances_} ).sort_values("importance", ascending=False)