Add standardization to the data processing pipeline

[tobmi1]

Right now, data normalization/standardization is performed speparately within each dataset class. Since the code is more or less the same everywhere, we should probably turn it into a Transform that can then be added to each dataset.