NumPy

#What you will learn

By the end of this tutorial you should be able to:

• explain why NumPy exists and when to reach for it instead of plain Python lists,
• create and inspect arrays (shape, dtype, ndim, size),
• index, slice, reshape and combine arrays,
• run arithmetic, statistical and boolean operations efficiently,
• handle missing data (NaN),
• save and load arrays to disk.

NumPy is the backbone of the Python data stack. Pandas, scikit-learn, PyTorch and TensorFlow all rely on it. Time spent here pays back in every data project that follows.

#Why NumPy?

Python is easy to read and write, but plain lists are not designed for numerical work. A list of one million numbers can occupy ten times more memory than the equivalent NumPy array, and a loop over a list is much slower than the same operation in NumPy.

NumPy (short for Numerical Python) solves this with the ndarray — an N-dimensional array stored as a contiguous block of memory of a single data type, with operations implemented in C.

The two key benefits are:

• Less memory. Each element occupies exactly the bytes its type needs (e.g. 2 bytes for int16), nothing else.
• More speed. Operations are vectorised — they run as a single C loop instead of a Python for loop.

#Setup

Create a new project with uv and add NumPy:

uv init numpy-lab
cd numpy-lab
uv add numpy

Open a Python REPL inside the project:

uv run python

By convention NumPy is imported as np:

import numpy as np

Every example in this tutorial assumes that import is already in place.

#Creating arrays

#From a Python list

The simplest way is to pass a list to np.array():

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