Software Development · Data Tooling · Open Source Library · 2026
FramePeek, a Lightweight EDA Library for Pandas
A Python package for fast, structured exploratory data analysis on pandas DataFrames. One function call produces a data overview, column profiling, missing value detection, duplicates, outliers, correlations, and data quality warnings. Built to remove repetitive boilerplate from the EDA process.
- 1 callEDA interfaceOne call for a structured profile
- 7 checksData quality profileOverview, columns, missing data, duplicates, outliers, correlations, and warnings
- 3 quality gatesCode quality controlspytest, mypy, and ruff
Case study contents
Objective
Build a lightweight, easy to use EDA tool so data scientists do not have to rewrite the same data checking code in every project.
Results and limitations
Produced a package that can be installed and used immediately for quick EDA on pandas DataFrames. Limitation, the check coverage is still basic and only supports pandas, with no support yet for other data libraries like polars or built in visualization.
Visual evidence
Technical details
Open implementation details
Role and contribution
Designed and built the library alone from scratch, including the package structure, core EDA functions, unit tests, and tooling setup like type checking and linting.
Methodology
Built a Python package with a src layout, designing a function that combines common data checks, such as column profiling, missing values, duplicates, outliers, and correlations, into a single function call. Added unit tests with pytest, type checking with mypy, and linting with ruff to maintain code quality.
Technologies
- Python
- pandas
- pytest
- mypy
- ruff
- setuptools