How it works...

The easiest way to perform operations on big datasets is to use NumPy's array class. As we've already seen, this class allows us to perform common operations, including basic mathematical operations between a scalar and an array. Converting the generator to an array, however, requires us to load the data into memory. Python's built-in list function takes an iterator and returns a list. This is required because the NumPy array must know the length of the data in order to allocate the correct amount of memory. With the array, it is easy enough to calculate the mean and then perform the divide equals scalar operation across the entire array. This broadcasts the division operation so that each element in the array is divided by the mean. We are, in essence, normalizing the elements of the array by their mean. We can then zip together our years with the newly computed data and return the time series.