Shapes: reshape and Stacking in Python

The same numbers can be laid out in different shapes. np.arange(12) gives back a flat array of the whole numbers 0 to 11, shape (12,). .reshape(rows, columns) gives back the same numbers as a table, filled row by row. Rows times columns must equal the count, or numpy raises a ValueError.

Write -1 for one size and numpy works it out: .reshape(-1, 1) means one column, as many rows as it takes. That is the shape scikit-learn wants for a single feature: one row per customer, one column for the feature. np.column_stack takes a list of flat arrays and gives back one 2-D array with each of them as a column.

ages = np.array([25, 41, 33])
kids = np.array([0, 2, 1])
grid = np.arange(8).reshape(2, 4)      # rows [0 1 2 3] and [4 5 6 7]
col = ages.reshape(-1, 1)              # shape (3, 1): 3 rows, 1 column
both = np.column_stack([ages, kids])   # rows [25 0], [41 2], [33 1]
print(both.shape)                      # (3, 2)

In this lesson, visits and spent hold the visits and spending of 4 customers. The starting code prints numbers and the shape of visits. When you're done, the program prints:

Numbers: [ 0  1  2  3  4  5  6  7  8  9 10 11]
[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]]
Visits shape: (4,)
[[ 3]
 [ 8]
 [ 1]
 [12]]
Column shape: (4, 1)
[[  3  40]
 [  8  95]
 [  1  12]
 [ 12 180]]

Your task: print numbers reshaped into 3 rows of 4, reshape visits into one column called column and print it with its shape, then stack visits and spent side by side with np.column_stack into table and print it.

Assignment
Follow the Coding Tutorial to reshape numbers into a grid, turn visits into a column and stack two arrays into a table.

Hint
.reshape gives back a new array and leaves the original as it was, so visits stays (4,). np.column_stack takes one argument, a list: put both arrays inside square brackets.