numpy

NumPy

NumPy is the foundational library for scientific computing in Python, providing multidimensional array objects, vectorized mathematical operations, and unified C/C++ API bindings.

When to Use

• High-Performance Numerical Computation: Vectorized array arithmetic and multi-dimensional tensor manipulation.
• Scientific Computing & Linear Algebra: Matrix decompositions (SVD, QR, Cholesky), eigenvalues, and inversions via np.linalg.
• Foundational Data Layer for AI/ML: Serving as the memory buffer standard for PyTorch, TensorFlow, OpenCV, and Scikit-learn.
• NumPy 2.0+ Modern Standard: Utilizing improved typing, standardized C-API, and optimized array constructors.

Quick Start

import numpy as np

# Vectorized matrix operations with broadcasting
a = np.array([[1, 2, 3], [4, 5, 6]], dtype=np.float64)
b = np.array([10, 20, 30])

# Broadcast addition across rows
c = a + b
print("Result shape:", c.shape)
print("Row-wise sum:", np.sum(c, axis=1))

Core Concepts

Vectorization, Broadcasting & Masking

Eliminating slow Python loops through vector operations:

import numpy as np

# Fast random number generation with modern Generator API
rng = np.random.default_rng(seed=42)

# Generate 2D array of sensor readings
readings = rng.normal(loc=25.0, scale=3.0, size=(1000, 5))

# Broadcasting: subtract sensor baseline offsets (shape (5,)) from matrix (1000, 5)
baselines = np.array([24.0, 25.0, 24.5, 25.5, 26.0])
calibrated = readings - baselines

# Boolean masking: identify readings exceeding alert threshold
anomaly_mask = np.abs(calibrated) > 6.0
anomaly_count = np.count_nonzero(anomaly_mask)

print(f"Total readings: {readings.size}, Anomalies detected: {anomaly_count}")

Linear Algebra & Matrix Decomposition with np.linalg

Solving linear equations and singular value decomposition:

# System of equations: Ax = b
A = np.array([[3.0, 1.0], [1.0, 2.0]])
b = np.array([9.0, 8.0])

# Solve for x: [x0, x1]
x = np.linalg.solve(A, b)
print("Solution x:", x)

# Singular Value Decomposition (SVD)
U, S, Vt = np.linalg.svd(A)
print("Singular values:", S)