Use tensorflow.keras (like the boston housing example)
  to fit a regression curve to the data in:
  c

  Learn about loading an .npz file with numpy (google it!)
  
  You will get three numpy arrays: x_train, y_train and x_test.
  Train a neural net and use it to predict the y's for x_test.
  Best (smallest) 'mae' wins.... a better grade+++.