Umap dimensionality reduction starter
A framework for discovering, compiling, and validating reusable skills for scientific agents.
npx -y skills add ma-compbio-lab/SkillFoundry --skill umap-dimensionality-reduction-starterAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
SKILL.md
0.8 KB, 161 tokens by cl100k_base, as published. Nobody here has run it
UMAP Dimensionality Reduction Starter
Use this skill to embed a tiny labeled feature matrix with UMAP and summarize the resulting 2D geometry.
What This Skill Does
- reads a toy feature matrix with labels
- computes a deterministic 2D UMAP embedding
- summarizes per-label centroids and separation
- emits machine-readable point coordinates for downstream plotting
Run
./slurm/envs/statistics/bin/python skills/statistical-and-machine-learning-foundations-for-science/umap-dimensionality-reduction-starter/scripts/run_umap_dimensionality_reduction.py --input skills/statistical-and-machine-learning-foundations-for-science/umap-dimensionality-reduction-starter/examples/toy_embedding_input.tsv --out scratch/statistics/umap_embedding_summary.json