pandas-data-cleaning

What problem does it solve?

Data cleaning is time-consuming and error-prone when done manually; this skill standardizes how to load, normalize, deduplicate, handle missing values, convert types, and save results for repeatable data pipelines.

Core Features & Use Cases

• Load data from CSV or Parquet and treat missing values with per-column strategies.
• Normalize column names to a consistent snake_case convention.
• Deduplicate rows using a natural key or composite key, with options to keep first or last.
• Convert data types to nullable dtypes (Int64, Float64, string) to preserve missing values.
• Save cleaned data to Parquet for fast I/O and dtype preservation.

Quick Start

Run a full cleaning workflow on a dataset by loading from CSV, normalizing columns, deduplicating on a natural key, handling missing values per dtype, converting types to nullable dtypes, and saving the result to Parquet.