Claude Skill MartBrowse skillsWhat is a Skill?
 
← Back to list

Exploratory Data Analysis

Runs bounded, network-free exploratory profiling of authorized local scientific files with missingness, leakage, and outlier diagnostics.

Data & AnalyticsIntermediate★ 33,030⑂ 3,248AI score 9/10Last updated: Aug 9, 2026

What it does

Inspects authorized local data files before modeling or confirmatory inference, producing deterministic aggregate reports.

  • Automated core: CSV/TSV schema and profile, missingness / group / split audits, distribution–outlier–transformation sensitivity, strict JSON structure checks
  • Optional inspectors: NumPy (.npy/.npz), HDF5 metadata, FASTA/FASTQ streaming summaries, PNG/JPEG/TIFF container metadata
  • Safety contract: no network calls, no pickle, raw data stays read-only, no raw rows/sequences/identifiers printed (tokenized instead), unknown formats fail closed
  • Report scaffold: generates a Markdown template for assumptions, sensitivity analyses, and limitations

It also enforces an EDA reasoning checklist — data dictionary, observational unit, batch/site structure, missing codes, train/val/test split basis — that helps prevent leakage and over-interpretation.

Who it's for

  • Researchers handling experimental or clinical data where reproducibility and audit trails matter
  • Data scientists who want to catch group/split leakage before training
  • Analysts in environments where identifiers and sensitive metadata must never be echoed
  • Teams juggling many scientific formats who need routing to validated domain tooling

Examples

  1. Clinical CSV screen: run missingness_leakage_audit.py with --entity-column subject_id --split-column split to see whether the same subject appears in both train and test.
  2. Distribution review: use distribution_sensitivity.py on a measurement column to compare mean/SD against median/IQR/MAD and inspect log-transform sensitivity without touching the data.
  3. Microscopy metadata check: run image_inspector.py on an OME-TIFF to read page/series/axes/dtype only, with no pixel decoding.

· · · Install guide · · ·

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/exploratory-data-analysis folder from the GitHub repo K-Dense-AI/scientific-agent-skills into my ~/.claude/skills/exploratory-data-analysis/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/K-Dense-AI/scientific-agent-skills.git /tmp/scientific-agent-skills && mkdir -p ~/.claude/skills && cp -r /tmp/scientific-agent-skills/skills/exploratory-data-analysis ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal.
  2. Clone the repository: git clone https://github.com/K-Dense-AI/scientific-agent-skills.git
  3. Create the skills directory: mkdir -p ~/.claude/skills
  4. Copy the skill: cp -r scientific-agent-skills/skills/exploratory-data-analysis ~/.claude/skills/
  5. Confirm Python 3.11+ is available: python3 --version (the CSV/TSV/JSON core uses only the standard library).
  6. For NumPy/HDF5/FASTA/image inspection, install Python 3.12+ and uv, then add only what you need: uv pip install numpy h5py biopython pillow tifffile (the pinned versions in SKILL.md are indicative — drop the pins if installation fails).
  7. Create a dedicated approved data directory (e.g. ~/approved-project); every command passes it via --root.
  8. Restart Claude Code and ask something like "run an EDA on this CSV" to trigger the skill.
View source on GitHub ↗License: MIT