NumPy — численные вычисления и работа с массивами

data

numpy is a Claude Code skill that establishes best practices for numerical computing with NumPy, covering array creation, broadcasting, vectorization, linear algebra, and memory management. It guides the use of explicit dtypes with constructors like `np.zeros`, `np.arange`, and `np.linspace`, replacing slow Python loops with vectorized operations and `np.where()` for conditional element-wise logic. Memory optimization techniques include choosing `np.float32` over `np.float64` when full precision is unnecessary, preferring views over copies, and using `np.memmap` for datasets that exceed available RAM. The skill also flags common pitfalls such as float equality with `==` instead of `np.allclose()`, silent integer overflow, and the deprecated `np.matrix`. Intended for data processing, scientific computing, and machine learning workflows where performance and memory efficiency matter.