Examples

The Zedda repository ships one example: a Jupyter notebook that walks through profiling, warning detection, and auto-cleaning on the Titanic dataset.

# Titanic Quickstart

File: examples/titanic_quickstart.ipynb

A 4-cell, 3-markdown-cell notebook. Loads the Titanic dataset from the public DataScienceDojo mirror, profiles it, lists warnings, and cleans it.

# Setup

pip install zedda[parquet]
jupyter lab examples/titanic_quickstart.ipynb

The [parquet] extra is required because the notebook passes a pandas DataFrame to Zedda, which writes a temporary Parquet file under the hood.

# Cell 1 — Load and profile

import pandas as pd
import zedda as zd

url = "https://raw.githubusercontent.com/datasciencedojo/datasets/master/titanic.csv"
df = pd.read_csv(url)

zd.profile(df)

This prints the full Rich EDA report: dataset overview panel, quality score, per-column table (Survived, Pclass, Sex, Age, SibSp, Parch, Ticket, Fare, Cabin, Embarked), smart warnings (high nulls on Age and Cabin, ID-like PassengerId, high-cardinality Name and Ticket), and any correlation alerts above |r| ≥ 0.7.

# Cell 2 — List every warning

zd.warnings(df)

Prints every issue ranked by severity (critical → warning → info), each with inline fix code, plus a copy-paste block at the end and the overall quality score.

# Cell 3 — Auto-clean

zd.clean(df, output="titanic_clean.csv")

The corrected call:

1. Writes titanic.csv.zedda-backup (or for DataFrame input, simply returns the cleaned DataFrame).
2. Applies every auto-fixable warning (drop PassengerId, Name, Ticket, Cabin; fill Age with median; etc.).
3. Writes titanic_clean.csv with a new zedda_source_file column.
4. Writes titanic_clean_cleaning_audit.json recording every fix applied.
5. Returns the cleaned pandas.DataFrame.

To restore the original:

zd.clean.undo("titanic.csv")

# More examples (in the docs)

The Guides section of this documentation has runnable snippets for every public function:

• Profiling Datasets — zd.profile() and zd.scan()
• Comparing Datasets — zd.compare()
• Cleaning & Fixing — zd.fix() and zd.clean()
• ML Readiness — zd.ml_ready()
• AI Q&A — zd.ask()
• HTML Reports — zd.report()

# Contributing an example

If you have a notebook or script that demonstrates a Zedda use case, contributions are welcome. See Contributing for the development setup.

# See also

• Quick Start — every public function in 30 seconds.
• Python API — every argument and return type.
• CLI — the command-line equivalents.