# geomstats/geomstats

Computations and statistics on manifolds with geometric structures.

Repository: https://github.com/geomstats/geomstats
Canonical: https://ross.abutalabs.com/products/geomstats
Homepage: http://geomstats.ai
Language: Python
License: MIT
License Family: permissive
Topics: statistics, geometry, manifold, machine-learning, deep-learning, gpu-programming, lie-groups, neural-networks, geodesic, riemannian-geometry
Last push: 2026-08-21T18:26:15+00:00

## Health v2 (maintenance only)
Score: 67/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 8, longevity 100
- inputs: {"age_days": 3235, "days_push": 12, "days_rel": 723, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1513, forks 295 (observed 2026-08-28T04:04:56.286954+00:00)

## What it is
Geomstats is an open-source Python package for computations, statistics, and machine learning on manifolds with geometric structures. It provides differential geometry operations (manifolds, Lie groups, Riemannian metrics) and geometric learning algorithms with a Scikit-Learn-like API, supporting NumPy, Autograd, and PyTorch backends.

## Use cases
- compute statistics on rotation matrices
- perform machine learning on manifold-structured data
- compute geodesics and Riemannian metrics
- cluster data lying on a Lie group like SO(3)
- learn geometric deep learning models with PyTorch backend
- analyze shape spaces of biological or natural shapes
- learn differential geometry hands-on with code examples

## When to choose
- your data lives on nonlinear manifolds such as rotations, SPD matrices, or shape spaces
- you need geometrically consistent statistics like Fréchet means instead of naive Euclidean averages
- you want a Scikit-Learn-style API for Riemannian learning algorithms
- you need to switch between NumPy, Autograd, and PyTorch backends

## When to avoid
- your data is standard Euclidean vectors and ordinary scikit-learn suffices
- you need a general-purpose deep learning framework rather than geometry-aware tools
- you are not comfortable with differential geometry concepts and have no need for them

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, data-science, math, simulation
- domain: machine-learning, mathematics, data-science, education
- platform: python, cross-platform
- tags: riemannian-geometry, manifold-learning, lie-groups, geometric-statistics, geodesics, differential-geometry, numpy, pytorch, scikit-learn-api, algorithms

## Member repositories
- geomstats/geomstats (main) score 67

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:56.286954+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:32:15.988345+00:00, confidence not recorded.
  - readme: https://github.com/geomstats/geomstats (fetched 2026-08-28T04:04:56.286954+00:00, sha 34dd02cd0f1b)
  - homepage: http://geomstats.ai (fetched 2026-08-29T11:36:01.543511+00:00, sha 0e18752ccbca)
  - registry_pypi: https://pypi.org/pypi/geomstats/json (fetched 2026-08-29T11:36:01.561694+00:00, sha 4eb69debc46f)
- Data as of 2026-08-30T08:39:29.467469+00:00.