Description

Generates per-column dtype, cardinality, null-rate, distribution summary and pairwise correlations. Designed to be readable in the REPL and exportable to HTML for review.

Parameters

Name	Type	Default	Description
data	str \| pathlib.Path \| pandas.DataFrame	—	Source dataset.
sample	int \| None	None	Randomly sample N rows before profiling for very large datasets.
correlations	bool	True	Compute pairwise Pearson correlations for numeric columns.

Returns

Profile — Object with .columns, .correlations, .to_html() and .to_dict().

Raises

ValueError — sample is larger than the dataset.

Examples

Profile a DataFrame

profile_example.py

python

Colab GitHub

import pandas as pd
import eazydatafix as edf

df = pd.read_csv("hospital.csv")
prof = edf.profile(df, sample=10_000)
prof.columns["age"]

Python 3.11— Expected output

>>> prof = edf.profile(df)
>>> prof.columns["age"]
ColumnProfile(age)
  dtype       int64
  missing     0
  unique      87
  min         0
  max         104
  mean        42.3

Notes

Profiles are lightweight and cheap to serialise — safe to store per-run.
Correlations only include columns with dtype numeric or bool.

Best Practices

Use sample= for datasets over ~5M rows to keep profiling under a second.
Export prof.to_html() into your CI artefacts for quick data reviews.

edf.profile()