numpy

import numpy as np
a = np.arange(12).reshape(3, 4)
print(a)
a.sum(axis=0), a.mean(), a[a % 3 == 0], a.T @ a

[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]]

(array([12, 15, 18, 21]), np.float64(5.5), array([0, 3, 6, 9]), array([[ 80,  92, 104, 116],
       [ 92, 107, 122, 137],
       [104, 122, 140, 158],
       [116, 137, 158, 179]]))

# Seeded streams match NumPy's exactly
np.random.seed(0)
x = np.random.randn(100_000)
x.mean(), x.std(), np.percentile(x, [5, 50, 95])

(np.float64(0.0015767005081253388), np.float64(0.9973396878273816), array([-1.64170792,  0.00172022,  1.644483  ]))

A = np.array([[4.0, 1, 2], [1, 3, 0], [2, 0, 5]])
w, v = np.linalg.eigh(A)
print(w)
np.linalg.solve(A, [1, 2, 3]), np.linalg.det(A), np.allclose(v @ np.diag(w) @ v.T, A)

[1.85489731 3.4760236  6.66907909]
(array([-0.30232558,  0.76744186,  0.72093023]), np.float64(43.000000000000014), True)

# Least squares and polynomials
x = np.linspace(0, 1, 50)
y = 3 * x**2 - x + 0.05 * np.sin(40 * x)
p = np.polyfit(x, y, 2)
p, np.roots(p)

(array([ 3.06605655, -1.06585058,  0.01298164]), array([0.33498999, 0.01263914]))

# A spectrum with the FFT
import matplotlib.pyplot as plt
t = np.linspace(0, 1, 512, endpoint=False)
signal = np.sin(2 * np.pi * 50 * t) + 0.5 * np.sin(2 * np.pi * 120 * t) + 0.3 * np.random.randn(512)
freq = np.fft.rfftfreq(512, d=1 / 512)
plt.figure(figsize=(7, 3)); plt.plot(freq, np.abs(np.fft.rfft(signal))); plt.xlabel('Hz'); plt.title('|FFT|')