All skills
oaustegard avatar

/exploring-data

@fde9b75

Exploratory data analysis. Use when users upload .csv/.xlsx/.json/.parquet files or request "explore data", "analyze dataset", "EDA", "profile data". Small files get ydata-profiling HTML/JSON reports; large files (over 200MB or 5M rows) get fixed-memory DuckDB/sketch profiling. Also covers near-duplicate row detection, cross-file key overlap ("can these join?"), dataset drift vs a stored baseline, and time-series profiling.

Use this Skill: https://skilld.dev/gh/oaustegard/claude-skills/exploring-data

This session only. Nothing lands on disk.

CHANGELOG.md

≈702 tokens on demand. Your agent reads this file only when SKILL.md points to it.

exploring-data - Changelog

All notable changes to the exploring-data skill are documented in this file. The format is based on Keep a Changelog.

[0.2.0] - 2026-09-06

Fixed

  • scripts/sketch_ops.py dups: build each row's MinHash from token counts, not the token set. MinHash.update keeps a per-permutation minimum, so hashing a token twice changed nothing and new york new york scored Jaccard 1.0 against york new. A token occurring c times now contributes c distinct elements (t, 0) … (t, c-1), which makes the signature estimate the weighted Jaccard index sum(min(a_i, b_i)) / sum(max(a_i, b_i)). Rows whose tokens are all distinct are unaffected.

Added

  • dups --unweighted: keeps the pre-0.2.0 set semantics for callers that want them.
  • tests/test_sketch_ops.py: covers both metrics, including a pair that differs only in token counts.

Changed

  • The dups summary line now names the metric it used — weighted Jaccard>= by default, Jaccard>= under --unweighted.

[0.1.2] - 2026-08-25

Fixed

  • repair broken frontmatter, mark obsolete skills, close registry gaps (#746)

Other

  • creating-skill: use Anthropic's quick_validate.py instead of a hand-rolled check (#775)
  • Deprecate mapping-codebases; adopt ruff 0.16.0 baseline (#747)

[0.1.1] - 2026-07-20

Fixed

  • install_ydata.sh: pin setuptools<81. ydata-profiling 4.18.x imports pkg_resources, removed in setuptools>=81; unpinned installs died on the first ProfileReport import with ModuleNotFoundError: pkg_resources.
  • install_ydata.sh: add pyarrow. analyze.sh reads .parquet via pandas, which needs a parquet engine that wasn't installed.

Added

  • scripts/visualize_findings.py: render an EDA profile as a compact, self-contained Chart.js dashboard (missingness tiers, most-skewed/zero-inflated distributions, largest categorical breakdowns) instead of only linking the dense ydata HTML. Reads both ydata JSON and profile_large.py --json output. Stdlib-only.
  • SKILL.md step A.5: present findings visually; prefer inline charts over a file link on surfaces that render visuals.

Added

  • add mapping-features skill for behavioral web app documentation (#432)
  • add line numbers, markdown ToC, and other files listing
  • add code maps and CLAUDE.md integration guidance
  • Delete VERSION files, complete migration to frontmatter
  • Migrate all 27 skills from VERSION files to frontmatter

Fixed

  • limit markdown ToC to h1/h2 headings only

Other

  • exploring-data v0.1.0: large-file profiling (DuckDB) + sketch ops (dups, overlap, drift) (#737)
  • exploring-data: use absolute path for check_install.sh in Step 2 (#710)
  • Remove _MAP.md files, direct agents to tree-sitting for code navigation (#545)

Source: SKILL.md on GitHub

3 warnings13d4 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    The skill provides robust exploratory data analysis capabilities using industry-standard libraries like ydata-profiling and DuckDB. It manages dependencies securely via virtual environments and includes logic for handling both small and large datasets. All external resources originate from trusted repositories and well-known services. No malicious patterns were detected.

  • Socket13d

    1 alert: gptAnomaly

  • Snyk13d

    Risk: MEDIUM · 1 issue

  • Runlayer7mo

    11/11 files flagged

Signed by skilld at fde9b75. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 4 weeks ago
metadata
{
  "version": "0.2.0"
}

README badge

README badge for oaustegard/claude-skills/exploring-data