Data Science & Numerical Computing Packages
Numerical computing, dataframes, and scientific-analysis libraries used across research and production data pipelines (NumPy, pandas, SciPy).
Packages in Data Science & Numerical Computing
Kind
Sort:Used by Most Apps
contourpy
Fast contour line and filled-polygon calculation for 2D grids
RealFFT
Fast real-to-complex forward FFT and complex-to-real inverse FFT for Rust, built on RustFFT.
downsample
Downsampling and smoothing algorithms for time series data, built for fast, accurate chart rendering.
jaxlib
The compiled XLA/PJRT runtime and native kernel library that powers JAX's array computation and program transformations.
marimo
A reactive Python notebook that's reproducible, git-friendly, and deployable as scripts or apps.
Bokeh
Interactive visualization library for building browser-based plots, dashboards, and data apps from Python.
prophet
Automatic forecasting library for time series with strong seasonality, using an interpretable additive model behind a scikit-learn-style API.
gensim
Topic modelling, document indexing, and similarity retrieval for Python.
Dask
A flexible parallel computing library that scales NumPy, pandas, and scikit-learn workflows across cores and clusters.
PyOD
The most comprehensive Python library for anomaly and outlier detection, with 60+ detectors across tabular, time series, graph, text, image, and audio data.
delta-spark
Python APIs for Delta Lake, the ACID storage layer for Apache Spark data lakes.
UMAP
A fast, scikit-learn-compatible dimension reduction library for visualizing and preprocessing high-dimensional data.
Ibis
A portable Python dataframe library with a lazy expression API across 20+ backends.
papermill
Parameterize, execute, and analyze Jupyter notebooks from the CLI or Python as reusable, repeatable pipeline steps.
torchdiffeq
GPU-accelerated, differentiable ODE solvers for PyTorch with O(1)-memory adjoint backpropagation for neural ODEs.
Rapier2d
Fast, cross-platform 2D physics engine for Rust games, animation, and robotics
gl-matrix
A JavaScript vector and matrix math library hand-tuned for high-performance WebGL and real-time 3D graphics applications.
nalgebra
General-purpose linear algebra library for the Rust ecosystem
TensorFlow Datasets
Ready-to-use, versioned public datasets exposed as tf.data.Dataset objects, with first-class support for TensorFlow, JAX, and NumPy pipelines.
AWS SDK for pandas
Pandas on AWS - integrate DataFrames with Athena, Glue, Redshift, S3, and more.
hdbscan
A high-performance, scikit-learn-compatible implementation of HDBSCAN for density-based clustering with automatic cluster count detection.
chDB
In-process OLAP SQL engine powered by ClickHouse, embedded directly in Python
ConnectorX
The fastest, most memory-efficient way to load data from databases into Python and Rust DataFrames.
tabula-py
Extract tables from PDF files directly into pandas DataFrames.
