# YingfanWang/PaCMAP

PaCMAP: Large-scale Dimension Reduction Technique Preserving Both Global and Local Structure

Repository: https://github.com/YingfanWang/PaCMAP
Canonical: https://ross.abutalabs.com/products/pacmap
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-07-20T23:56:21+00:00

## Health v2 (maintenance only)
Score: 87/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 93, release rhythm 72, longevity 100
- inputs: {"age_days": 2216, "days_push": 44, "days_rel": 184, "gap_med": 6, "n_releases_24m": 6}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1016, forks 85 (observed 2026-08-28T04:03:14.436194+00:00)

## What it is
PaCMAP is a Python library implementing Pairwise Controlled Manifold Approximation, a dimensionality reduction technique that preserves both local and global structure of high-dimensional data for visualization. It is the official implementation of the JMLR-published method, with related extensions like Parametric PaCMAP and LocalMAP.

## Use cases
- visualize high-dimensional embeddings in 2D
- reduce dimensionality of single-cell or embedding datasets
- find a UMAP/t-SNE alternative that preserves global structure
- plot clusters in large datasets
- explore manifold structure of feature vectors

## When to choose
- you need 2D/3D visualizations balancing local and global structure
- you want a fast, well-tested Python dimension reduction library
- you are comparing manifold learning methods for research

## When to avoid
- you need supervised dimensionality reduction with labels
- you need GPU parametric online-learning DR (use ParamRepulsor instead)
- you need exact linear projections like PCA

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, data-visualization, math
- domain: machine-learning, data-visualization, data-science
- platform: python, cross-platform
- tags: dimensionality-reduction, manifold-learning, embedding, umap-alternative, visualization

## Member repositories
- YingfanWang/PaCMAP (main) score 87

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:14.436194+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 2026-08-30T07:11:17.147436+00:00, confidence not recorded.
  - readme: https://github.com/YingfanWang/PaCMAP (fetched 2026-08-28T04:03:14.436194+00:00, sha 9093c354c813)
  - registry_pypi: https://pypi.org/pypi/pacmap/json (fetched 2026-08-29T13:10:16.740400+00:00, sha 5110d332aeb7)
- Data as of 2026-08-30T08:39:29.467469+00:00.