YData Profiling

4.18.1 · active · verified Sat Apr 11

ydata-profiling is a powerful Python library that automates the generation of comprehensive exploratory data analysis (EDA) reports for pandas DataFrames. It provides detailed statistics, visualizations, and interactive widgets to understand data quality and distributions. The library is actively maintained with frequent minor releases, typically monthly or bi-monthly.

Warnings

breaking The library was renamed from `pandas-profiling` to `ydata-profiling` starting with version 3.0.0. This change requires updating import statements and package names in `requirements.txt` or `pyproject.toml`.
Fix: Update `pip install pandas-profiling` to `pip install ydata-profiling` and `from pandas_profiling import ...` to `from ydata_profiling import ...`.
gotcha Profiling large datasets can be computationally intensive and consume significant memory. For very large datasets, consider sampling or using the Spark integration (`ydata-profiling[spark]`).
Fix: For large datasets, use `df.sample()` or enable the Spark backend. Monitor memory and CPU usage during report generation. Tune the `config_minimal` or `config_small` parameters for faster, less detailed reports.
gotcha Specific features like Jupyter Notebook widgets or Spark DataFrame profiling require installing optional dependencies. The base `pip install ydata-profiling` does not include these.
Fix: Install with extras: `pip install ydata-profiling[notebook]` for widgets or `pip install ydata-profiling[spark]` for Spark integration.
gotcha The library has specific Python version requirements (currently `Python >=3.10, <3.14`). Ensure your environment matches these requirements to avoid installation or runtime errors.
Fix: Check your Python version using `python --version` and upgrade or downgrade as necessary, or use a virtual environment with a compatible Python version.

Install

pip install ydata-profiling Base installation
pip install ydata-profiling[notebook] For Jupyter Notebook widgets
pip install ydata-profiling[spark] For Spark DataFrame support

Imports

ProfileReport
```
from ydata_profiling import ProfileReport
```
The library was renamed from 'pandas-profiling' to 'ydata-profiling' in version 3.0.0. The import path changed accordingly.

Quickstart

This quickstart demonstrates how to create a pandas DataFrame, generate a comprehensive profile report using `ProfileReport`, and save it to an HTML file. For interactive display in Jupyter Notebooks, use `profile.to_widgets()`.

import pandas as pd
from ydata_profiling import ProfileReport

# Sample DataFrame
data = {
    'col1': [1, 2, 3, 4, 5],
    'col2': ['A', 'B', 'A', 'C', 'B'],
    'col3': [1.1, 2.2, None, 4.4, 5.5]
}
df = pd.DataFrame(data)

# Generate the profile report
profile = ProfileReport(df, title="My DataFrame Profile", explorative=True)

# Save report to an HTML file
profile.to_file("your_report.html")

# If running in a Jupyter Notebook, you can display widgets directly:
# profile.to_widgets()

print("Profile report saved to your_report.html")

view raw JSON →