Contributing

The dataset repository is privately managed by maintainers. All contributions flow through the UI Form — no repository access needed.

Before You Begin

  • Read Data Spec — required fields, validation rules, dropdown values

  • Read Schema Reference — standard table and column names

  • Read Examples — see what good entries look like

  • Base your use case on a real or realistic business scenario from any domain

Submission Steps

Step 1 — Open the UI Form

Open the Query Entry Form in your browser. No login or account needed.

Step 2 — Fill All Required Fields

Complete all 8 sections. The Live JSON panel updates in real time. Required sections: Meta, Business Question, Business Context, Metrics & Aggregation, Schema Tables, Data Model Layers. Optional: Chain of Thought, SQL Answer.

Step 3 — Copy the JSON

Copy the full JSON output from the Live JSON panel on the right side of the form.

Step 4 — Submit to Maintainers

Paste your JSON in a new Discussion thread with the title format:

[Submission] <domain>_<topic>_<short_description>

Maintainers will review and ingest into the dataset.

Contribution Guidelines

Do:

  • Use real-world or realistic business scenarios from any supported domain

  • Write instruction as a business user would naturally ask it

  • Write aggregation_logic in plain language — avoid SQL jargon

  • Use standard schema table names from Schema Reference

  • Include SQL if possible — significantly improves dataset quality

  • Make sure required_metrics_kpis list matches keys in aggregation_logic

Don’t:

  • Include sensitive data, real customer names, or PII

  • Reference internal systems or proprietary company names

  • Use non-standard table names without noting the variation

  • Submit incomplete entries with empty required fields

  • Submit duplicates already covered in Examples

Review Process

  1. You post JSON in a Discussion thread

  2. Maintainer reviews for quality, correctness, completeness

  3. Approved → ingested into dataset, thread marked resolved

  4. Changes needed → feedback left in the thread

  5. Not suitable → declined with reason

Target review time: 7 days.

Quality Checklist

Before submitting verify:

  • [ ] All required sections filled

  • [ ] required_metrics_kpis matches keys in aggregation_logic

  • [ ] data_model.facts and data_model.dims non-empty

  • [ ] SQL (if provided) runs against standard schema without errors

  • [ ] No proprietary, internal, or PII data included