DataFrames & Scientific Python

pandas

pandas-dev/pandas

Popular

The ubiquitous Python DataFrame library for data cleaning, analysis and I/O.

NumPy

numpy/numpy

Popular

Fundamental N-dimensional array library that nearly all scientific Python builds on.

Polars

pola-rs/polars

Trending

Fast multi-threaded DataFrame library in Rust with a lazy query engine; the main pandas alternative.

Apache Arrow

apache/arrow

Popular

Standard columnar in-memory format and libraries (PyArrow) enabling zero-copy data exchange between tools.

SciPy

scipy/scipy

Popular

Core algorithms for optimization, statistics, signal processing and linear algebra in Python.

Ibis

ibis-project/ibis

Portable Python DataFrame API that compiles to SQL for DuckDB, BigQuery, Snowflake and 20+ backends.

Narwhals

narwhals-dev/narwhals

Rising

Lightweight compatibility layer that lets libraries support pandas, Polars, PyArrow and more with one API.

cuDF

NVIDIA/cudf

NVIDIA RAPIDS GPU DataFrame library that can accelerate pandas and Polars code with no changes.

statsmodels

statsmodels/statsmodels

Statistical modeling, hypothesis tests and econometrics with R-style formulas in Python.