dfhealth 0.1.2

Opinionated data health checks for pandas DataFrames

dfhealth 🏥

Stop shipping dirty data. dfhealth is an opinionated, zero-config tool to sanity-check your pandas DataFrames before they hit your models.

Think of it like PyLint for DataFrames.

Why dfhealth?

Data scientists often spend 80% of their time cleaning data. dfhealth automates the "sanity check" phase by flagging silent killers in your dataset with actionable advice.

Key Checks

Code	Name	What it catches
D001	Duplicate Rows	Identical rows that might skew your statistics.
I001	ID Integrity	Duplicate values in columns that look like IDs (e.g., user_id).
N001	Negative Values	Negative numbers in fields meant to be positive (e.g., age, price).
O001	Extreme Outliers	Values outside the 3x IQR range that could be data entry errors.
T001	Date-like String	Object columns that look like they should be datetime objects.

📦 Installation

pip install dfhealth

Quick Start

import pandas as pd
import dfhealth as dh

# Load your dataset
df = pd.read_csv("data.csv")

# Run the health check
report = dh.health_check(df)

# Get a beautiful, colored report in your terminal
report.print()

📖 License

MIT License - see the LICENSE file for details.

🤝 Contributing

Want to add a new check? Check out CONTRIBUTING.md.