Scientific Computing with Python - NumPy | WeiYuan

Scientific Computing with Python
- NumPy
2017/08/03 (Thus.)
WeiYuan

site: v123582.github.io
line: weiwei63
§ 全端⼯程師 + 資料科學家
略懂⼀點網站前後端開發技術，學過資料探勘與機器
學習的⽪⽑。平時熱愛參與技術社群聚會及貢獻開源
程式的樂趣。

Outline
§ About NumPy
§ Ndarray
§ Create a new Array
§ Property of Array
§ Operation of Array
§ Matrix
§ Advanced Usages
3

Outline
§ About NumPy
§ Ndarray
§ Matrix
§ Advanced Usages
4

the Ecosystem of Python
5Reference: https://www.edureka.co/blog/why-you-should-choose-python-for-big-data

6Reference: https://www.slideshare.net/gabrielspmoreira/python-for-data-science-python-brasil-11-2015

About NumPy
§ NumPy is the fundamental package for scientific computing
with Python. It contains among other things:
• a powerful N-dimensional array object
• sophisticated (broadcasting) functions
• tools for integrating C/C++ and Fortran code
• useful linear algebra, Fourier transform, and random number
capabilities
• be used as an efficient multi-dimensional container of generic data.
7

About NumPy
§ NumPy is the fundamental package for scientific computing
with Python. It contains among other things:
• a powerful N-dimensional array object
• sophisticated (broadcasting) functions
• tools for integrating C/C++ and Fortran code
• useful linear algebra, Fourier transform, and random number
capabilities
• be used as an efficient multi-dimensional container of generic data.
8

Try it!
§ #練習：Import the numpy package under the name np
9

Try it!
§ #練習：Print the numpy version and the configuration
10

Outline
§ About NumPy
§ Ndarray
§ Matrix
§ Advanced Usages
11

Ndarray
§ shape
§ ndim
§ dtype
§ size
§ itemsize
§ data
12
1
2
3
4
5
6
7
8
9
10
11
12
from numpy import *
a = array([[ 0, 1, 2, 3, 4],
[ 5, 6, 7, 8, 9],
[10, 11, 12, 13, 14]
])

Ndarray
§ shape
§ ndim
§ dtype
§ size
§ itemsize
§ data
13
1
2
3
4
5
6
7
8
9
10
11
12
from numpy import *
a = array([[ 0, 1, 2, 3, 4],
[ 5, 6, 7, 8, 9],
[10, 11, 12, 13, 14]
])
a.shape # (3, 5)
a.ndim # 2
a.dtype.name # 'int32’
a.size # 15
a.itemsize # 4
type(a) # numpy.ndarray

Outline
§ About NumPy
§ Ndarray
§ Matrix
§ Advanced Usages
14

Create a new Ndarray
§ One Dimension Array (1dnarray)
§ Multiple Dimension Array (ndarray)
§ Zeros, Ones, Empty
§ arange and linspace
§ random array
§ array from list/tuple
15

16
1
2
3
4
5
6
7
8
9
import numpy as np
arr1 = np.array([0, 1, 2, 3, 4]) # array([0 1 2 3 4])
type(arr1) # <type 'numpy.ndarray'>
arr1.dtype # dtype('int64')
array( )

17
1
2
3
4
5
6
7
8
9
import numpy as np
arr1 = np.array([0, 1, 2, 3, 4]) # array([0 1 2 3 4])
arr1.dtype # dtype('int64')
arr2 = np.array([1.2, 2.4, 3.6]) # array([1.2, 2.4, 3.6])
arr2.dtype # dtype('float64')
array( )

§ Question：How to assign data type for an array ?
18

§ Multiple Dimension Array (ndarray)
19
1
2
3
4
5
6
7
8
9
import numpy as np
a = np.array([[1, 2, 3], [4, 5, 6]])
# array([[1, 2, 3],
# [4, 5, 6]])
array( )

§ Question：How to change shape from 1-d array ?
20

21
1
2
3
4
5
6
7
8
9
import numpy as np
zeros = np.zeros(5)
# array([ 0., 0., 0., 0., 0.])
zeros( )

22
1
2
3
4
5
6
7
8
9
import numpy as np
zeros = np.ones(5)
# array([ 1., 1., 1., 1., 1.])
ones( )

23
1
2
3
4
5
6
7
8
9
import numpy as np
zeros = np.empty(5)
# array([ 0., 0., 0., 0., 0.])
empty( )

24
1
2
3
4
5
6
7
8
9
import numpy as np
arange = np.arange(5)
# array([0 1 2 3 4])
arange( )

25
1
2
3
4
5
6
7
8
9
import numpy as np
linspace = np.linspace(0, 4, 5)
# array([ 0., 1., 2., 3., 4.])
linspace( )

§ random array
26
1
2
3
4
5
6
7
8
9
import numpy as np
linspace = np.random.randint(0, 2, size=4)
# array([ 0, 1, 1, 1])
random.randint( )

Try it!
§ #練習：Create a 3x3x3 array with random values
27

Try it!
§ #練習：Find indices of non-zero elements
28

§ array from list/tuple
29
1
2
3
4
5
6
7
8
9
import numpy as np
x = [1,2,3]
a = np.asarray(x)
x = (1,2,3)
a = np.asarray(x)
asarray( )

Try it!
§ #練習：Create a null vector of size 10
30

Try it!
§ #練習：Create a vector with values ranging from 10 to 49
31

Try it!
§ #練習：Create a 3x3 identity matrix
32

Outline
§ About NumPy
§ Ndarray
§ Matrix
§ Advanced Usages
33

Property of Ndarray
§ shape
§ ndim
§ dtype
§ size
§ itemsize
§ data
34
1
2
3
4
5
6
7
8
9
10
11
12
import numpy as np
a = np.array(
[[11, 12, 13, 14, 15],
[16, 17, 18, 19, 20],
[21, 22, 23, 24, 25],
[26, 27, 28 ,29, 30],
[31, 32, 33, 34, 35]
])

Property of Ndarray
§ shape
§ ndim
§ dtype
§ size
§ itemsize
§ data
35
1
2
3
4
5
6
7
8
9
10
11
12
print(type(a))
print(a.shape)
print(a.ndim)
print(a.dtype)
print(a.size)
print(a.itemsize)
print(a.nbytes)

Try it!
§ #練習：How to find the memory size of any array
36

data type
37
§ Question：How to assign data type for an array ?
1. Set dtype with create the array
2. Change dtype function

1. Set dtype with create the array
38
1
2
3
4
5
6
7
8
9
x = numpy.array([1,2.6,3], dtype = numpy.int64)
print(x) #
print(x.dtype) #
x = numpy.array([1,2,3], dtype = numpy.float64)
print(x) #
print(x.dtype) #
array( )

2. Change dtype function
39
1
2
3
4
5
6
7
8
9
x = numpy.array([1,2.6,3], dtype = numpy.float64)
y = x.astype(numpy.int32)
print(y) # [1 2 3]
print(y.dtype)
z = y.astype(numpy.float64)
print(z) # [ 1. 2. 3.]
print(z.dtype)
astype( )

data shape
41
§ Question：How to change shape from 1-d array ?
1. Set multiple array with create the array
2. Assign new shape to shape property
3. Change shape function

1. Set multiple array with create the array
42
1
2
3
4
5
6
7
8
9
import numpy as np
a = np.array([[1,2,3],[4,5,6]])
a.shape # (2, 3)
array( )

2. Assign new shape to shape property
43
1
2
3
4
5
6
7
8
9
a = np.array([[1,2,3],[4,5,6]])
a.shape = (3,2)
a # [[1, 2], [3, 4], [5, 6]]
array( )

44
1
2
3
4
5
6
7
8
9
a = np.array([[1,2,3],[4,5,6]])
b = a.reshape(3,2)
b # [[1, 2], [3, 4], [5, 6]]
reshape( )

45
1
2
3
4
5
6
7
8
9
a = np.array([[1,2,3],[4,5,6]])
b = a.reshape(3,2)
b # [[1, 2], [3, 4], [5, 6]]
a.resize(3,2)
a # [[1, 2], [3, 4], [5, 6]]
resize( )

Try it!
§ #練習：Create a 3x3 matrix with values ranging from 0 to 8
46

index and slicing
§ index
§ slicing
47
1
2
3
4
array[0] # 0
array[1] # 1
array[-1] # 4
1
2
3
4
5
array[1:3] # [1, 2]
array[:4] # [0, 1, 3]
array[3:] # [3, 4]
array[1:4:2] # [1, 3]
array[::-1] # [4, 3, 2, 1, 0]
([0, 1, 2, 3, 4])

index and slicing
§ index
§ slicing
48
1
2
3
4
array[1] # [0, 1]
array[1][0] # 0
array[1][1] # 1
array[2][0] # 2
1
2
3
4
5
array[0:2] # [[0, 1], [2, 3]]
array[:2] # [[0, 1], [2, 3]]
array[2:] # [[4, 5]]
( [0, 1, 0, 1, 0],
[2, 3, 2, 3, 2],
[4, 5, 4, 5, 4] )

index and slicing
§ slicing
49
1
2
3
4
array[0, 1:4] # [1, 0, 1]
array[[0, 0, 0], [1, 2, 3]]
array[1:3, 0] # [1, 3, 5]
array[[1, 2], [0, 0, 0]]
( [0, 1, 0, 1, 0],
[2, 3, 2, 3, 2],
[4, 5, 4, 5, 4] )

Try it!
§ #練習：Create a null vector of size 10 but the fifth value which
is 1
50

Try it!
§ #練習：Reverse a vector (first element becomes last)
51

Try it!
§ #練習：Create a 2d array with 1 on the border and 0 inside
52

Try it!
§ #練習：Create a 8x8 matrix and fill it with a checkerboard
pattern
53

Try it!
§ #練習：
54
1
2
3
4
5
6
7
8
9
import numpy as np
a = np.array([[1,2,3],[3,4,5],[4,5,6]])
# 第二列元素
# 第二行元素
# 除了第二列的元素

Try it!
§ #練習：
55
1
2
3
4
array[::2,::2]
array[:, 1]

Outline
§ About NumPy
§ Ndarray
§ Matrix
§ Advanced Usages
56

sophisticated (broadcasting)
58

Outline
§ About NumPy
§ Ndarray
§ Matrix
§ Advanced Usages
61

Outline
§ About NumPy
§ Ndarray
§ Matrix
§ Advanced Usages
65

Advanced Usages
§ Boolean indexing and Fancy indexing
§ Boolean masking
§ Incomplete Indexing
§ Where function
§ Customize dtype
66

Thanks for listening.
2017/08/03 (Thus.) Scientific Computing with Python – NumPy
Wei-Yuan Chang
v123582@gmail.com
v123582.github.io

Scientific Computing with Python - NumPy | WeiYuan

Scientific Computing with Python - NumPy | WeiYuan

More Related Content

What's hot

Similar to Scientific Computing with Python - NumPy | WeiYuan

More from Wei-Yuan Chang

Recently uploaded

In this document

Scientific Computing with Python - NumPy | WeiYuan