4.1 NumPy arrays

Big ideas

• import numpy as np. An array holds many numbers of one type and does math on all of them at once: a * 2, a + b, a ** 2 (element-wise, no loop).
• Making arrays: np.array([1, 2, 3]), np.arange(0, 10, 2), np.linspace(0, 1, 5), np.zeros(4), np.ones((2, 3)), np.random.randint(1, 7, size=10).
• a.shape, a.size, a.dtype. Index and slice like lists; 2-D arrays use a[row, col], a[:, 0] for a column.
• Boolean masks filter: a[a > 5]; (a > 5).sum() counts. Reductions: a.sum(), a.mean(), a.max(), a.argmax(), a.std(), with axis=0 down columns or axis=1 across rows.
• Arrays are the fast, memory-light way to hold data; pandas and matplotlib are built on them.

Code you should know

import numpy as np

temps = np.array([61, 68, 75, 80, 72])
print(temps * 2)
print((temps - 32) * 5 / 9)
print(temps.mean(), temps.max(), temps.argmax())
print(temps[temps > 70], (temps > 70).sum())
print(np.arange(0, 10, 3), np.linspace(0, 1, 5))

[122 136 150 160 144]
[16.11 20.   23.89 26.67 22.22]
71.2 80 3
[75 80 72] 3
[0 3 6 9] [0.   0.25 0.5  0.75 1.  ]

import numpy as np

grid = np.array([[1, 2, 3], [4, 5, 6]])
print(grid.shape, grid.size, grid[1, 2], grid[:, 0])
print(grid.sum(axis=0), grid.sum(axis=1))
print(np.zeros((2, 2)), np.ones(3))

(2, 3) 6 6 [1 4]
[5 7 9] [ 6 15]
[[0. 0.]
 [0. 0.]] [1. 1. 1.]

Worked example

Common mistakes

• Mixing a Python list and an array: [1, 2] * 2 repeats the list; np.array([1, 2]) * 2 doubles the numbers.
• Off-by-one with arange (stop excluded) versus linspace (stop included).
• Forgetting axis and getting a single number when you wanted per-column results.