# Data Reviewer

> Use for read-only checks of economics datasets, data maps, variable coding, panel balance, sample restrictions, and synthetic-data caveats before estimation or replication.

- Skill: `meleantonio/data-reviewer` (Agent Skill)
- Install (CLI): `npx skillmds@latest add meleantonio/data-reviewer`
- Raw SKILL.md: https://api.skillmd.com/api/skills/meleantonio/data-reviewer/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: meleantonio (https://skillmd.com/u/meleantonio)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/meleantonio/data-reviewer

---


# Data Reviewer

Audit whether the data are understandable, documented, and safe to use.

## Checklist

1. Read the README, data source map, and analysis script before inspecting data.
2. Confirm the data source is named and the bundled/public/private status is clear.
3. Confirm the unit of observation and expected panel keys.
4. Confirm treatment, comparison group, wave/timing, and outcome coding.
5. Check missing values and impossible values for analysis variables.
6. Check panel balance or document why balance is not required.
7. Confirm sample restrictions and transformations are documented.
8. Confirm synthetic or teaching data are clearly labelled where relevant.
9. Return blockers first; do not edit files unless separately instructed.

## Card-Krueger Checks

For `examples/card-krueger/`, verify:

- `state` is only `NJ` or `PA`;
- `wave` is only `before` or `after`;
- `fte_employment` is numeric and non-negative;
- each `store_id` has one before row and one after row;
- `docs/data_source_map.md` says the CSV is synthetic teaching data.

## Output

```markdown
# Data Review

| Check | Status | Evidence |
| --- | --- | --- |

## Blockers

## Documentation Gaps

## Non-Blocking Suggestions
```