Zero-Boilerplate ML Data Validation That Actually Feels Good to Use

NEO LINE — Sat, 25 Jul 2026 13:30:27 +0000

I love data validation, but I hate writing 500 lines of YAML just to check a CSV. Tools like Great Expectations are powerful, but they feel way too heavy for quick checks, internal scripts, or fast-moving projects.

So I built data-fitcheck — a lightweight Python library that does three things without any config:

Validates CSVs (missing data, outliers, type mismatches)
Detects data drift between training and production sets
Evaluates sklearn models and generates self-contained HTML reports

30-Second Quickstart


bash
pip install data-fitcheck && fitcheck demo
GitHub: https://github.com/neoline361-art/fitcheck
PyPI: https://pypi.org/project/data-fitcheck/

I built a zero-boilerplate alternative to Great Expectations for ML data validation

NEO LINE — Sat, 25 Jul 2026 10:30:34 +0000

I love data validation, but I hate writing 500 lines of YAML just to check a CSV. So I built data-fitcheck — a lightweight Python library that does three things without any config:

Validates CSVs (missing data, outliers, type mismatches)
Detects data drift between training and production sets
Evaluates sklearn models and generates self-contained HTML reports

Install it:
bash

pip install data-fitcheck

Run the demo:
bash

fitcheck demo

It instantly generates three HTML reports in your folder. No setup, no YAML, no database.

GitHub: https://github.com/neoline361-art/fitcheck
PyPI: https://pypi.org/project/data-fitcheck/

I'd love feedback on the API design or feature requests!

DEV Community: NEO LINE

Zero-Boilerplate ML Data Validation That Actually Feels Good to Use

30-Second Quickstart

I built a zero-boilerplate alternative to Great Expectations for ML data validation