Answer 1

Because machine learning algorithms and deep learning neural networks require that input and output variables be numbers, they do not know what to do with text, labels, strings or categorical data... This means that categorical data must be encoded to numbers before we can use it to fit and evaluate a model