pandas & Polars

DataFrames are how market data is shaped in practice: timestamps on the index, instruments as columns or rows, and operations expressed as selections, joins, and groupbys. pandas remains the lingua franca — label-based indexing, automatic alignment on timestamps, resampling irregular ticks into bars, and split-apply-combine with groupby. Polars is the newer challenger: a multi-threaded, query-optimized engine with a lazy API that often runs the same transformation an order of magnitude faster on large datasets and with clearer out-of-memory behavior. Learn pandas first for its ecosystem, then reach for Polars when data gets big, and keep both built on the array thinking from the previous node.