# lmcinnes/umap

Uniform Manifold Approximation and Projection

Repository: https://github.com/lmcinnes/umap
Canonical: https://ross.abutalabs.com/products/umap
Homepage: https://umap-learn.readthedocs.io
Language: Python
License: BSD-3-Clause
License Family: permissive
Topics: umap, dimensionality-reduction, visualization, machine-learning, topological-data-analysis
Last push: 2026-08-21T10:17:06+00:00

## Health v2 (maintenance only)
Score: 87/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 66, longevity 100
- inputs: {"age_days": 3350, "days_push": 12, "days_rel": 147, "gap_med": 31, "n_releases_24m": 10}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 8255, forks 868 (observed 2026-08-28T04:10:19.852769+00:00)

## What it is
UMAP is a Python library implementing Uniform Manifold Approximation and Projection, a non-linear dimensionality reduction technique usable for visualization and general embedding. It models data with a fuzzy topological structure and finds low-dimensional projections that preserve it, serving as a faster alternative to t-SNE.

## Use cases
- visualize high-dimensional data in 2D or 3D
- reduce dimensions before clustering
- find a faster t-SNE alternative for large datasets
- embed single-cell gene expression data
- transform new data points into an existing embedding
- preprocess features for machine learning pipelines

## When to choose
- you need fast, scalable non-linear dimension reduction
- you want cluster-preserving 2D/3D visualizations of high-dimensional data
- you need to embed new data into an existing embedding or use inverse transforms

## When to avoid
- you need strictly linear dimensionality reduction like PCA
- you require exact preservation of global distances rather than local structure
- you need a pure interpretable statistical model rather than an embedding

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, data-visualization, data-science
- domain: machine-learning, data-science, data-visualization
- platform: python, windows
- tags: dimensionality-reduction, umap, topological-data-analysis, manifold-learning, embedding, tsne-alternative, linux, macos

## Member repositories
- lmcinnes/umap (main) score 87

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:19.852769+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 2026-08-29T17:29:28.058819+00:00, confidence not recorded.
  - readme: https://github.com/lmcinnes/umap (fetched 2026-08-28T04:10:19.852769+00:00, sha c06816614eac)
  - homepage: https://umap-learn.readthedocs.io (fetched 2026-08-29T08:27:54.853900+00:00, sha aabc8fa7fc6d)
- Data as of 2026-08-30T08:39:29.467469+00:00.