Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Dataset Curation Methodology

A research-grade playbook for diagnosing dataset bias and imbalance, building stratified splits, planning expansion, and running an ethics review.

Data & AnalyticsIntermediate★ 368⑂ 31AI score 7/10Last updated: Jul 3, 2026

What it does

Guides Claude through a 6-step dataset curation workflow.

  • Distribution analysis: per-class counts, imbalance ratio (max/min), rare-class detection (<5% of largest class), label co-occurrence matrix, spurious correlation checks
  • Bias assessment: three screening questions (real-world reflective? harmful? fixable?) plus fairness metrics such as demographic parity, equalized odds, and representation ratio
  • Stratified sampling: primary stratification by label, secondary by source to prevent leakage, chi-squared validation, split ratios by dataset size (80/10/10 for large, k-fold for <5k)
  • Quality assessment: inter-annotator agreement via Cohen's/Fleiss' kappa or Krippendorff's alpha, label noise estimation, edge-case discovery
  • Expansion plan: priority classes, source suggestions, active learning / targeted scraping / synthetic augmentation, cost estimates
  • Ethics checklist: sensitivity, consent, privacy, licensing, misuse potential, datasheet/data card documentation

Who it's for

  • Grad students and researchers assembling or releasing a dataset for a paper
  • ML engineers whose models underperform on specific subgroups due to skewed training data
  • Data teams that must report labeling quality and agreement metrics quantitatively

Examples

  1. "Analyze the class imbalance in this image dataset" → get per-class counts, imbalance ratios, rare-class list, and co-occurrence findings.
  2. "Split into train/val/test without letting clips from the same movie cross splits" → get a label+source dual stratification strategy with chi-squared validation.
  3. "Run an ethical review before I publish this dataset" → get a completed checklist on consent, privacy, licensing, and misuse risk plus data-card guidance.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/fcakyon/phd-skills/HEAD/plugin/skills/dataset-curation/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

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 plugin/skills/dataset-curation folder from the GitHub repo fcakyon/phd-skills into my ~/.claude/skills/dataset-curation/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/fcakyon/phd-skills.git && mkdir -p ~/.claude/skills && cp -r phd-skills/plugin/skills/dataset-curation ~/.claude/skills/

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

  1. Open a terminal (Terminal on macOS/Linux, Git Bash on Windows).
  2. Clone the repo: git clone https://github.com/fcakyon/phd-skills.git
  3. Create the skills folder: mkdir -p ~/.claude/skills
  4. Copy the skill in: cp -r phd-skills/plugin/skills/dataset-curation ~/.claude/skills/
  5. Verify with ls ~/.claude/skills/dataset-curation — you should see SKILL.md.
  6. Restart Claude Code and say something like "check my dataset for class imbalance" to trigger it.
View source on GitHub ↗License: MIT