Documentation
EazyDataFix 2.0 is a deterministic-first Python framework for reusable recipes, reviewed transformations and Analysis Ready, ML Ready and Power BI Ready workflows.
Installation
EazyDataFix is on PyPI and requires Python 3.10 or later (tested with Python 3.10–3.13). Install with pip in the environment of your choice — a virtualenv is recommended.
$ pip install eazydatafixQuick Start
Start with a PipelineRecipe when preparation must be reviewed and reused, or run edf.run() for the complete profile → assess → clean → EDA workflow. See the quick start guide for a full walkthrough.
Core APIs
The stable public API spans data quality, preparation, validation and deterministic Agentic EDA:
edf.PipelineRecipeandedf.ColumnRule— versioned JSON/YAML preparation rules with explicit types and failure policies.edf.Pipeline— preview, apply or deliberately run a recipe with transformation evidence and guarded exports.edf.run()— unified profile, assessment, controlled cleaning and EDA.edf.assess()— dataset quality report.edf.fix()— opinionated auto-cleaning pipeline.edf.profile()— column-level profiling.edf.prepare_with_report()— deterministic preparation with changes and warnings.edf.analysis_ready_with_report()— analysis-ready data with before/after evidence.edf.ml_ready()— leakage-safe train/test inputs and reusable preprocessing.edf.powerbi_ready()— validated single-table or multi-table Power BI inputs.edf.infer_schema()andedf.validate_contract()— pipeline data contracts.edf.run_agentic_eda()— end-to-end deterministic Agentic EDA.
Examples
The Examples gallery walks through seven realistic datasets — CSV, Excel, HR, healthcare and more.
Roadmap
v2.0.0 is the current stable release. It adds reusable preparation recipes, explicit conversions, review plans, quarantine evidence and a shared pipeline interface while preserving the v1 convenience APIs. Read the full roadmap.