Data Analysis

Why vectorization and broadcasting make NumPy fast, how pandas builds a labeled Series/DataFrame/Index model on top of raw arrays, and where Polars' lazy query-plan model breaks from both - the mental model behind every other page in this section.

The NumPy/pandas mental model and a first analysis.

Creation, dtypes, broadcasting, and vectorized math.

Slicing, fancy indexing, and axes.

Construction, selection, and indexing.

Missing values, dtypes, and string ops.

Split-apply-combine and pivot tables.

Combining datasets without silent duplication.

Resampling, rolling windows, and timezones.

Reproducible, vectorized, well-typed analysis.