Missing Data in Scikit-learn Models

In what ways do models in scikit-learn manage missing data? What strategies can you, as a data scientist, implement to address missing values effectively?