Now we've covered the basics of reading and writing dataset objects; we can begin transforming our loaded dataset. When reading from file, these operations are not needed as they are built-in to read functions like tf.data.experimental.make_csv_dataset It takes nothing more than this single line to shuffle and batch our dataset! We can use the map function to perform operations on each sample within xour dataset. For example, for predicting the next time-step in a sequence, we may want to train on input data, which consists of timesteps n to n 8, and output data consisting of timesteps n 1 to n 9. Initially, our dataset may consist of many samples containing sequences of 10 time-steps.