NumPy is the core Python library for numerical computing, centred on the ndarray object that stores homogeneous data in contiguous memory. It supplies vectorised operations, broadcasting, linear algebra, Fourier transforms, and random number generation.
The official site brings together installation instructions for common platforms, beginner and advanced tutorials, a searchable API reference, release notes, and guides for migrating between major versions, plus links to the broader scientific Python ecosystem.
It suits data analysts cleaning and reshaping tabular data, researchers implementing simulations or statistical methods, and engineers building preprocessing pipelines that feed machine learning frameworks or visualisation tools.
Compared with pure Python lists, NumPy arrays are far faster and more memory efficient for bulk arithmetic; compared with dataframe libraries it works at a lower level, so labelled columns and grouped operations often require an additional layer such as pandas.
Pin a version when reproducing results, watch for silent dtype and shape changes during upgrades, and prefer vectorised expressions over Python loops; consult the migration notes before moving across major releases.