Introduction to Numpy

Table of Contents

In data analysis, machine learning, and scientific computing, efficiently handling data is essential. Numpy is a simple and powerful Python library that provides efficient array computation. This article introduces Numpy, including basic syntax, attributes, and common operations.

Introduction
#

The main data structure in Numpy is the array. Arrays are similar to Python lists, with one key difference: a Numpy array can only store elements of the same type, unlike lists which can contain mixed types. Numpy also provides many mathematical functions, such as computing the median, mean, variance, and more.

Creating Arrays in Numpy
#

To use Numpy, we must first import it—typically with the alias np:

import numpy as np

The simplest way to create an array is to convert a list using np.array:

arr1 = np.array([1, 2, 3])
arr2 = np.array([[1, 2, 3], [4, 5, 6]])  # a 2D array

Other Methods to Create Arrays
#

np.zeros(n): Creates an array of length n filled with zeros. You can also specify a shape, such as np.zeros((2, 3)) for a 2×3 array.
np.ones(n): Creates an array of length n filled with ones.
np.arange(start, end, step): Creates an array like [start, start + step, start + 2*step, ...]. The final element is less than end, so end is not included.

Array Attributes
#

Numpy arrays have some useful attributes:

ndim: Returns the number of dimensions.
shape: Returns the shape of the array. For example, arr2 is a 2×3 array, so it returns (2, 3).
dtype: Returns the data type, such as int64.
itemsize: Tells you how many bytes each element occupies.

Common Operations
#

Concatenation
#

np.concatenate is used to combine arrays:

arr1 = np.array([1, 2, 3])
arr2 = np.array([4, 5, 6])
arr3 = np.array([7, 8, 9])
np.concatenate((arr1, arr2, arr3), axis=0)

Here, axis=0 means stacking them row-wise. Since arr1, arr2, and arr3 all have shape (3,), the final result is a (9,) array: [1, 2, 3, ..., 9]. Note: You cannot concatenate along axis=1 unless the arrays are reshaped.

Reshape
#

reshape changes the shape of an array. For example, to convert arr1 to shape 1×3:

arr1 = arr1.reshape(1, 3)

To stack arrays side by side like this:

[[1 2 3]
 [4 5 6]
 [7 8 9]]

You need to reshape each array to 3×1, then concatenate along axis=1:

arr1 = arr1.reshape(3, 1)
arr2 = arr2.reshape(3, 1)
arr3 = arr3.reshape(3, 1)

result = np.concatenate((arr1, arr2, arr3), axis=1)

Sorting
#

Numpy offers two ways to sort an array:

np.sort(arr): Returns a sorted copy of the array. The original array remains unchanged.
arr.sort(): Sorts the array in-place, modifying the original array.

Indexing
#

In addition to positive indices, you can use negative indices to access elements from the end. For example, arr[-1] returns the last element, arr[-2] returns the second-to-last, etc.

You can also use slicing to get subarrays, such as arr[1:4] to get elements with indices 1 through 3 (the end index is exclusive).

Lastly, you can use conditional filtering to get elements that match certain criteria. For example: