import numpy as np

data = np.array([[1, 1.5, 2], [5, 3, 1]])

data.shape

data.dtype

data.ndim

np.zeros((3, 6))

np.empty((2, 4))

np.arange(10)

np.arange(10, dtype = 'float64')

np.arange(10 == 'float64').dtype

a1 = np.array([1, 2, 3, 4.5], dtype=np.float64) a1

a1.dtype

# you can convert data type of an array. #note that the fractional part is turnicated a2 = a1.astype(np.int64) a2

a1 = np.array([1, 2, 3, 4])

a1 * a1

1 / a1

a1 ** 4

a2 = np.array([4, 3, 2, 1]) a1 > a2

a1 = np.arange(10) a1

# changing index 2, 3, 4 (not 5) to 2000 a1[2: 5] = 2000

# a1 slice is a1 in subset index 2, 3, 4 a1s = a1[2:5] a1s

# remember, a1s is not a new list, new variable, # its just a subsect of a1, a1s is a1 but dealing, # with a small group of a1 and anything affecting, # a1s will affect a1 a1

a1s[1] = 10 a1s

a1

a3d = np.array([[[1, 2, 3], [4, 5, 6]], [[7, 8, 9], [10, 11, 12]]])

a3d

# 0 chooses which super list (the first list) # 1 chooses which sub list (the second list) # 2 chooses which number in the list a3d[0][1][2]

a3d[0, 1, 2]

a3d[0] = 5 a3d

a2d = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])

a2d

a2d[:2]

a2d[:2, 1:]

a = np.arange(18).reshape(3, 6) a

a.T # T is transpose

a.swapaxes(0,1)

np.dot(a.T, a)

a = np.arange(5) a

np.sqrt(a)

np.exp(a)

b = np.arange(4, -1, -1) b # start, stop, steps, start at 4, stop at -1, in increments of subtracting 1

c = a / 3 c

# r can be whatever, its just a place holder #modf calculates intigral fractions i , r = np.modf(c) i

r

a = np.arange(5) +1 a

b = np.arange(5) + 6 b

condition = np.array([True, False, True, True, False]) condition

# input x if c (condition) is true, else (false), input y [(x if c else y) for x, y, c in zip(a, b, condition)]

from numpy import random as rng a = rng.standard_normal((4, 4)) a

a > 0

np.where(a > 0, 1, -1)

a = np.arange(12).reshape(3,4) + 1 a

a.mean(axis = 1)

a.mean(axis = 0)

a.cumsum(axis = 0)

names = np.array(["Bob", "Will", "Joe", "Bob", "will", "Joe", "Joe"]) names

np.unique(names)

sorted(set(names))