Introduction to NumPy: The Backbone of Numerical Computing in Python

In the world of data science, machine learning, and scientific computing, NumPy stands as one of the most fundamental and essential libraries in Python. NumPy, short for Numerical Python, provides a powerful N-dimensional array object and a collection of routines for fast operations on arrays, including mathematical, logical, shape manipulation, sorting, selecting, and more.

Why NumPy?

• Efficient: NumPy is highly efficient in handling large amounts of data, providing functionalities that perform computations faster than native Python code.

• Array Operations: It allows you to perform element-wise operations on arrays, similar to matrix operations.

• Integration: Integrates well with other libraries like SciPy, Pandas, and Matplotlib, making it versatile for various types of data manipulation and visualization tasks.

Getting Started with NumPy

First, you need to install NumPy. You can install it using pip:

pip install numpy

Once installed,you can import it into your python script or Jupyter notebook:

import numpy as np

NumPy Array

The core of NumPy is the ndarray(N-dimension array) object.Unlike Python list, NumPy arrrays are homogenius,meaning all elements have to be of the same type. This homogeneity allows NumPy to efficiently perform operations over array.

Creating Array

you can create arrays in multiple ways:
1.From Lists:

import numpy as np
my_list= [1,2,3,4,5]
my_array=np.array(my_list)
print(my_array)

Output : [1 2 3 4 5 ]

2.Using built-in function:

zeros_array = np.zeros(5) #Create an array of zeros
ones_array=np.ones((2,3)) #Create a 2*3 array of ones
arange_array = np.arrange(10) #Create an array with values from 0 to 9

3.Random Array:

random_array= np.random.random(3,3) #create a 3*3 array with random value

Array Attributes

NumPy arrays come with several attributes that help you understand the array:

array = np.array([[1, 2, 3], [4, 5, 6]])

print("Shape:", array.shape)  # (2, 3)
print("Data Type:", array.dtype)  # int64 (or int32 depending on your system)
print("Size:", array.size)  # 6
print("Number of Dimensions:", array.ndim)  # 2

Array Operations

NumPy allows various operations on arrays. Here are some essential ones:

Element-wise Operations

You can perform arithmetic operations on arrays just like you would with scalars:

a = np.array([1, 2, 3])
b = np.array([4, 5, 6])

print("Addition:", a + b)  # [5 7 9]
print("Subtraction:", a - b)  # [-3 -3 -3]
print("Multiplication:", a * b)  # [4 10 18]
print("Division:", a / b)  # [0.25 0.4 0.5]

Aggregate Functions

NumPy provides several functions to perform statistical operations on arrays:

array = np.array([1, 2, 3, 4, 5])

print("Mean:", np.mean(array))  # 3.0
print("Sum:", np.sum(array))  # 15
print("Standard Deviation:", np.std(array))  # 1.4142135623730951
print("Maximum:", np.max(array))  # 5
print("Minimum:", np.min(array))  # 1

Matrix Operations

NumPy supports matrix operations, which are crucial in data science and machine learning:

matrix1 = np.array([[1, 2], [3, 4]])
matrix2 = np.array([[5, 6], [7, 8]])

print("Matrix Multiplication:\n", np.dot(matrix1, matrix2))

Output:

[[19 22]
 [43 50]]

Reshaping and Broadcasting

Reshaping arrays allows you to change their dimensions without altering the data:

array = np.arange(12)
reshaped_array = array.reshape((3, 4))
print(reshaped_array)

Output:

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

Broadcasting lets you perform operations on arrays of different shapes:

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

print("Broadcasted Addition:", array + scalar)  # [3 4 5]

Advanced feature

Fancy Indexing and Slicing:

NumPy provides powerful indexing and slicing capabilities to access subsets of data:

array = np.array([1, 2, 3, 4, 5])

# Slicing
print("Slice:", array[1:4])  # [2 3 4]

# Fancy Indexing
indices = [0, 2, 4]
print("Fancy Indexing:", array[indices])  # [1 3 5]

Boolean Masking

You can use boolean conditions to filter data:

array = np.array([1, 2, 3, 4, 5])

mask = array > 2
print("Boolean Masking:", array[mask])  # [3 4 5]

Linear Algebra Functions

NumPy includes a submodule dedicated to linear algebra operations:

from numpy.linalg import inv, eig

matrix = np.array([[1, 2], [3, 4]])
inverse = inv(matrix)
eigenvalues, eigenvectors = eig(matrix)

print("Inverse:\n", inverse)
print("Eigenvalues:\n", eigenvalues)
print("Eigenvectors:\n", eigenvectors)

Whether you're handling large datasets, performing complex mathematical operations, or simply want faster computations, NumPy is the go-to library.
Investing time in mastering NumPy will undoubtedly pay off as you dive deeper into the realms of data analysis and machine learning. Its robust features and versatility will enhance your productivity and open up new possibilities for your projects.

np.rand and np.randn

A number of functions are available in NumPy's random module to create ndarrays initialized with random values. For example, here is a 3x4 matrix initialized with random floats between 0 and 1 (uniform distribution):

np.random.rand(3,4)

Output :

array([[ 0.37454012,  0.95071431,  0.73199394,  0.59865848],
       [ 0.15601864,  0.15599452,  0.05808361,  0.86617615],
       [ 0.60111501,  0.70807258,  0.02058449,  0.96990985]])

np.arange

You can create an ndarray using NumPy's range function, which is similar to python's built-in range function:

np.arange(1, 5)

Output :

array([1, 2, 3, 4])

Reshaping an array

In place

Changing the shape of an ndarray is as simple as setting its shape attribute. However, the array's size must remain the same.

g = np.arange(24)
print(g)
print("Range:",g.dim)

Output :

[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23]
Rank: 1

g.shape=(6,4)
print(g)
print("Rank:",g.ndim)

Output :

[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]
 [12 13 14 15]
 [16 17 18 19]
 [20 21 22 23]]
Rank: 2

reshape

The reshape function returns a new ndarray object pointing at the same data. This means that modifying one array will also modify the other.

g2 = g.reshape(4,6)
print(g2)
print("Rank:", g2.ndim)

Output :

[[ 0  1  2  3  4  5]
 [ 6  7  8  9 10 11]
 [12 13 14 15 16 17]
 [18 19 20 21 22 23]]
Rank: 2

Thanks for reading ~ Ujjwal Paliwal