Vectorization is an important concept in Natural Language Processing (NLP) and is used to represent text data as vectors. These vectors are then projected in an N-dimensional plane to capture similarities and patterns in the data. This helps Machine Learning algorithms to learn patterns and make predictions.