Added Percentify just above polars and under pandas

[dmitriy1ikobe]

Adding Percentify → Data Analysis.
Percentify

A data analysis toolkit for exploring and profiling datasets quickly.

Acceptance criterion: Hidden Gem.

304 stars, 44 forks — within the preferred 100–500 range.
~Almost a year old (created Sept 2025) with consistent activity — releases through v0.1.0 to v1.0.2 with active commits this month both from owner and contributors.
On PyPI: pip install percentify (v1.0.2).

Why it's different;
Percentify prides itself on being a one-line-import data analysis toolkit. It implements multicollinearity analysis using VIF, an important technique that is buried deep inside "statsmodels.stats.outliers_influence.variance_inflation_factor".

It makes important data analysis tools like VIF available through a single import and a simple, consistent interface, meaning users do not need to know where each technique is implemented across different statistical libraries.

It also contains a profiler, but unlike YData Profiling, it leans more towards diagnostics than description. It profiles the data immediately, says what is wrong with it, prioritises the problems, and gives recommendations on what to do. This is especially important for learners getting into data analysis, because it does not only present statistics; it also helps users understand what those statistics mean and what action they can take.

The profiler returns structured, programmatic results rather than only producing a visual report. Its findings can therefore be filtered, exported, tested, or included in automated data-analysis and data-quality workflows.

It also includes functions such as outlier detection, permutation tests, effect sizes, confidence intervals, PCA, class-imbalance checks, and other tools that are commonly needed during exploratory data analysis.

What I love about the project, and what I think makes it unique, is the fact that it supports both pandas and Polars through the same public interface. Users can apply the same functions without having to learn or maintain separate implementations for each DataFrame library.

It also implements informative warnings instead of direct errors for recoverable issues. These warnings can help learners understand questionable inputs or analytical limitations, while more experienced programmers can turn them off when necessary or during production.

Basically, the package covers what is important for quick data analysis while making it reachable, actionable, and teachable.