AI Skill MartBrowse skillsPromptsProgramsLearnWhat is a Skill?
 
← Back to list

Dataset Curation Methodology

A step-by-step methodology for analyzing dataset distribution and bias, building stratified splits, planning expansion, and running an ethics review.

Data & AnalyticsIntermediate★ 446⑂ 38AI score 6/10Last updated: Nov 1, 2026

What it does

  • Distribution analysis: per-class counts, imbalance ratio (max/min), detection of severely underrepresented classes (<5% of max), label co-occurrence matrix
  • Bias assessment: spurious correlation detection plus fairness metrics such as demographic parity, equalized odds, and representation ratio
  • Stratified splitting: primary stratification by label, secondary by source to prevent leakage, chi-squared validation, and size-based split ratios (80/10/10, 70/15/15, or k-fold)
  • Annotation quality: Cohen's kappa, Fleiss' kappa, Krippendorff's alpha, label-noise estimation, edge-case review
  • Expansion & ethics: priority classes, collection strategies (active learning, targeted scraping, synthetic augmentation), and a consent/privacy/licensing/misuse checklist

Who it's for

  • ML engineers and researchers who build or clean their own training datasets
  • Anyone writing a fairness report, datasheet, or data card
  • Teams doing an ethics review before publishing a public dataset

Examples

  1. "Analyze class imbalance in my labeled data" → class count table, imbalance ratios, and a list of underrepresented classes
  2. "Split train/val/test without source leakage" → dual stratification strategy with chi-squared distribution checks and zero source overlap
  3. "Run an ethical review before I release this dataset" → completed checklist covering sensitivity, consent, anonymization, licensing, and misuse risk

· · · 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/labrat/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 AI Skill Mart.
Copy the plugin/skills/dataset-curation folder from the GitHub repo fcakyon/labrat into my ~/.claude/skills/fcakyon-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/labrat.git && mkdir -p ~/.claude/skills && cp -r labrat/plugin/skills/dataset-curation ~/.claude/skills/

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

  1. Open a terminal.
  2. Create the skills directory: mkdir -p ~/.claude/skills
  3. Clone the repo: git clone https://github.com/fcakyon/labrat.git
  4. Copy just this skill: cp -r labrat/plugin/skills/dataset-curation ~/.claude/skills/
  5. Restart Claude Code and try prompts like "dataset bias analysis" or "create a stratified sample".
  6. Install pandas and scikit-learn in your analysis environment, since the skill ships no computation scripts of its own.
View source on GitHub ↗License: MIT