NumPy: Python's Mathematical Backbone

Welcome to the fascinating world of NumPy! If you’re new to data science or just starting to dip your toes into coding, you’ve landed in the perfect place.

NumPy, short for Numerical Python, is a fundamental package for scientific computing in Python. It’s the backbone that supports a wide array of data science operations, from basic mathematical operations to complex machine learning algorithms. In this guide, we’ll break down the basics of NumPy in a simple, engaging manner, peppered with code examples to ensure you grasp the concepts thoroughly.

What is NumPy?

NumPy is an open-source Python library that provides support for large, multi-dimensional arrays and matrices, along with a collection of mathematical functions to operate on these arrays. It’s designed for scientific computation, enabling complex mathematical operations to be expressed in a few lines of code. At its core, the library’s main object is the powerful ndarray (n-dimensional array), which is a grid of values, all of the same type, indexed by a tuple of non-negative integers.

Main Features

NumPy Library is used to work with arrays.
It also has functions for matrices, linear algebra, and Fourier Transform.
Python has lists which do the work of an array, but NumPy will provide 50x more speed than a Python list.
NumPy is written in C, C++ and Python.
Numpy integrates c and C++ codes in Python.
Where fast computations are required c, C++ is used.
It works efficiently with the latest CPU architecture.

Why Use NumPy?

Before NumPy, Python was capable of handling lists. However, lists are slow to process and not suitable for mathematical operations, especially when dealing with large datasets. NumPy arrays are stored at one continuous place in memory, unlike lists, making it efficient for:

Operations involving mathematical computations on arrays
Handling large data sets
Implementing machine learning algorithms

Numpy Installation

Using this command will ensure the installation of the latest version of NumPy that is compatible with your system.

pip install numpy

Importing Numpy into your code

import numpy

import numpy as np

Python

import numpy

num_arr=numpy.array([42,541,234,4,45,454,5])

print(num_arr)

print(type(num_arr))

Output

[ 42 541 234   4  45 454  5]
<class 'numpy.ndarray'>

NumPy Arrays: The Basics

Creating a NumPy array is straightforward. Let’s explore a few ways to do this.

From a Python List

Python

# Create a 1D NumPy array 
my_array = np.array([1, 2, 3, 4, 5]) 
print(my_array) 
# Create a 2D NumPy array (also known as a matrix) 
my_2d_array = np.array([[1, 2, 3], [4, 5, 6]]) 
print(my_2d_array)

Special Arrays in NumPy

NumPy also provides functions to create arrays filled with zeros, ones, or a range of numbers – handy for initializing large arrays.

Python

import numpy

num_arr=numpy.array([42,541,234,4,45,454,5])

print(num_arr)

print(type(num_arr))

Output

[ 42 541 234   4  45 454  5]
<class 'numpy.ndarray'>

# Array of zeros 
zeros = np.zeros((3, 4)) 
print(zeros) 
# Array of ones 
ones = np.ones((2, 2)) 
print(ones) 
# Array with a sequence of numbers 
sequence = np.arange(10, 20) 
print(sequence) 
# Evenly spaced numbers over a specified interval 
linspace = np.linspace(0, 1, 5) 
print(linspace)

Operations with NumPy Arrays

One of the strengths of NumPy is its ability to perform operations on arrays with ease.

Arithmetic Operations

You can easily perform arithmetic operations on arrays element-wise.

Python

a = np.array([1, 2, 3, 4]) 
b = np.array([10, 20, 30, 40]) 
# Addition 
print(a + b) 
# Subtraction 
print(b - a) 
# Multiplication 
print(a * b) 
# Division 
print(b / a)

Mathematical Functions

NumPy offers a vast library of mathematical functions to perform operations such as trigonometric calculations, logarithms, and more.

Python

import numpy

num_arr=numpy.array([42,541,234,4,45,454,5])

print(num_arr)

print(type(num_arr))

Output

[ 42 541 234   4  45 454  5]
<class 'numpy.ndarray'>

# Square roots 
print(np.sqrt(a)) 
# Exponential 
print(np.exp(a)) 
# Sine function 
print(np.sin(a))

Aggregations

Performing statistical calculations is straightforward with NumPy’s aggregation functions.

Python

# Sum 
print(np.sum(a)) 
# Maximum value 
print(np.max(a)) 
# Mean 
print(np.mean(a))

Numpy array Functions

Functions	Explanation
copy()	Copy functions create an identical, independent array from the original.
view()	The view reflects the original array, and any change to the view or the array is similarly affected.
shape()	Returns the shape of the array with respect to its dimensions.
reshape()	Changes the dimensions or changes the number of elements in the array.
concatnate()	Concatenate joins two or more arrays by using its 0th axis default and the axis is changeable.
stack()	Stack joins two 1D arrays into one using axis parameters.
hstack()	H Stack helps to stack arrays with the rows.
vstack()	V stack helps to stack arrays with the columns.
dstack()	D stack helps to stack arrays with height.
array_split()	split an array into a specific number of arrays.
hsplit()	split 2D arrays among their rows into a specific number of arrays.
where()	searches an element in the array and returns the indexes of matches.
searchsorted()	SearchSorted() is used for inserting elements according to order in the array and Returns an index for the element to be inserted.
sort()	Returns a sorted array into a specified order.
numpy.broadcast	numpy. broadcast

Python list vs Numpy array

Numpy array	Python List
A numpy array is limited to a single datatype in single array.	A numpy array is limited to a single datatype in a single array.
The Numpy array occupies less memory.	Python list calculations are slower than Numpy array
Numpy arrays are at the heart of all Numpy systems.	Python lists occupy more memory while containing similar items.
Numpy array calculations are faster because of the above fact.	Python has multiple sequenced data storage techniques that have different functionalities.
you can use ndarray to create multidimensional arrays. etc matrices , data tables.	Python has nested data set capabilities.

Numpy Ecosystem

Numpy has created a vast ecosystem spanning numerous fields of science.

It empowers the users with the computational powers of the C and FORTRAN languages. All these functionalities have been brought over to Python. Since the codes are easily readable, they can be used to make complex calculations. Numpy provides the ability to exploit highly functional machines’ powers for calculations. So many libraries exploit these capabilities to provide their functions. Such as Tensorflow, xtensor, Dask, CuPy, JAX, and Xarray.

In the field of data science, Numpy is capable of assisting at all steps of analytics.

ETL analysis
Exploratory analysis
Model creation and Evaluation
Dashboard reports

This article only introduces the basic, fundamental features of numpy. In the following articles, you will be introduced to a plethora of numpy features.

Numpy Arrays

Numpy Library is used to work with arrays. It also has functions for matrices, linear algebra, and Fourier Transform. Python has lists which do the work of an array, but numpy will provide 50x more speed than a Python list.

Distinctive Analytics