Feature scaling is often required when working with machine learning algorithms. This is because many machine learning algorithms require that the data is in a certain range in order for them to work properly. For example, if you are working with a neural network, the data needs to be scaled between 0 and 1 in order for the algorithm to converge.